The Great Ordovician Biodiversification Event (The Critical Moments and Perspectives in Earth History and Paleobiology)

The Great Ordovician Biodiversification Event

Critical Moments and Perspectives in Earth History and Paleobiology

1

Introduction Barry D. Webby

The historical scientist focuses on detailed particulars—one funny thing after another—because their coordination and comparison permits us, by consilience of induction, to explain the past with . . . confidence (if the evidence is good). stephen jay gould, Wonderful Life

T

his introductory chapter presents the scope, aims, and organization of the volume; outlines previous work on Ordovician biodiversity topics; and gives an overview of the chapters in the volume, from those briefly appraising the Ordovician World to those more comprehensively surveying the diversification patterns of the main Ordovician taxonomic groups. The chapter ends with some closing remarks on the Ordovician Radiation and future directions. ■

Scope and Aims

Two of the greatest evolutionary events in the history of life on earth occurred during Early Paleozoic time. The first was the Cambrian explosion of skeletonized marine animals—what Wilson (1992: 188) called the “big bang of animal evolution”— about 540 million years (m.y.) ago. The second was the “Great Ordovician Biodiversification Event,” the focus of this book. During the 46-m.y. span of Ordovician time (489–443 m.y. ago), an extraordinarily varied range of evolutionary radiations of “Cambrian-, Paleozoic- and Modern-type” biotas appeared, the

most diversified occurring in marine continental platform to open ocean habitats. Animals (arthropods) also first walked on land, and based on their cryptospore record, the first nonvascular bryophytelike plants colonized damp areas on land. The Cambrian explosion of skeletonized animals is now comparatively well documented, though the timing of the initial event still needs to be reassessed (Bowring and Erwin 1998). Books have been published on a number of aspects of this explosion, for example, Glaessner’s The Dawn of Animal Life: A Biohistorical Study (1984), Lipps and Signor’s Origin and Early Evolution of the Metazoa (1992), and Zhuravlev and Riding’s The Ecology of the Cambrian Radiation (2000). Also well documented is the history of the extraordinary Mid Cambrian Burgess Shale faunas, about 505 m.y. ago. The most notable publications are Gould’s Wonderful Life (1989), Briggs, Erwin, and Collier’s The Fossils of the Burgess Shale (1994), and Conway Morris’s The Crucible of Creation (1998b). In contrast, until very recently, there was no Ordovician biodiversity volume, or at least none that focused significantly on aspects of Ordovician biodiversity. 1

 .  1,000

A

Additional families with poorly mineralized skeletons

800

Number of Families

600

400

Md 200

0

Pz Cm –C

V

O

S D

600

P TR

C

400

J

K

T

200

0

Geologic Time (Millions of Years) 2,000

B

1,600

Modern EF

1,200

800

Paleozoic EF

400

Cambrian EF 0

U

Cambrian 510

T 1a–d

Time Slices

Ar

Ln L 4a–c

Lower Middle

489

C 5a–d

As Ly

Upper

Silurian

Ordovician 472

460.5

W Lv

6a–c

M

2a–c 3a–b

However, nearly half the contributions in Crame and Owen’s Palaeobiogeography and Biodiversity Change: The Ordovician and Mesozoic-Cenozoic Radiations (2002) have addressed issues relating to Ordovician diversity change. The important role of plate tectonics during the Ordovician was singled out—examples, such as the fragmentation of the Gondwanan margin, the drift of Avalonia, and the development of an array of marginal to oceanic terranes (including island chains) that effectively partitioned ocean circulation patterns within the Iapetus Ocean—as prominent in engendering the patterns of diversity change. Ordovician biodiversity has typically been studied in three different ways: (1) taxonomic diversity, which involves a focus on the taxonomic richness and turnover (originations and extinctions) within fossil groups; (2) ecologic diversity, which examines how organisms (or their communities) adapt to fill niche spaces in order to exploit available food resources more successfully; and (3) morphological diversity (usually termed disparity), which traces the patterns of morphological (design) change in various fossil groups. This book is devoted primarily to documenting the taxonomic diversification of Ordovician biotas, in both global and regional contexts, and to firming up the timing of the most important diversification events. The volume has a broad coverage, but inevitably there are gaps. A few aspects of ecologic diversity have been treated, but the coverage is limited, and aspects of morphological diversity (disparity) are barely touched on. Moreover, only limited discussion of possible causes for the radiation events is included. All these Ordovician diversification events form part of the “Great Ordovician Diversification Event,” also called the Ordovician Radiation (Droser et al. 1996). Although the most intensive part of the Ordovician Radiation was during the Mid to Late Ordovician epochs, an interval of 28 m.y. (until the second extinction pulse of the end Ordovician mass extinctions—Sheehan 2001b; chapter 9 in this volume), some taxonomic groups (e.g., trilobites, inarticulated brachiopods, graptolites, conodonts, and rostroconch mollusks) also diversified significantly during the Early Ordovician. Consequently, all these evolutionary events from the beginning to virtually the end of the Ordovician Period—through nearly 46 m.y. of earth history—should be treated as belonging to the Ordovician Radiation (figure 1.1).

Number of Genera

2

443

419 Ma

FIGURE 1.1. (A) Phanerozoic taxonomic diversity of marine animal families (slightly modified from Sepkoski 1984: figure 1). The fields Cm, Pz, and Md represent the Cambrian, Paleozoic, and Modern “evolutionary faunas” (EFs), respectively. Note also the generalized field of additional, poorly preserved families at the top of the diagram. Sepkoski’s timescale follows Harland et al. (1982) and includes the following abbreviations: V = Vendian; –C = Cambrian; O = Ordovician; S = Silurian; D = Devonian; C = Carboniferous; P = Permian; TR = Triassic; J = Jurassic; K = Cretaceous; T = Tertiary. (B) The Middle to Upper Cambrian, Ordovician, and Silurian (except Pridoli) taxonomic diversity of marine animal genera (modified from Sepkoski 1995: figure 1). The main field of Cambrian, Paleozoic, and Modern EFs is shown, as well as Sepkoski’s time units, including British series for his Ordovician and Silurian subdivisions. His abbreviations are as follows: M = Middle Cambrian; U = Upper Cambrian; T = Tremadocian; Ar = Arenig; Ln = Llanvirn; L = “Llandeilo”; C = Caradoc; As = Ashgill; Ly = Llandovery; W = Wenlock; Lv = Ludlow. The “Llandeilo” series has now been abandoned in favor of an enlarged Llanvirn, the name Llandeilian retained for its upper stage (figure 2.1), and the overlying Caradoc expanded downward to fill the gap (Fortey et al. 1995, 2000). Note also the added scale bar at the bottom of the diagram that comprises the tripartite global series subdivisions for the Ordovician System, radiometric dates in millions of years (Ma), and the time-slice subdivisions used in this volume.

Introduction No comparable, well-defined major extinction episode has been recognized at the beginning of the Ordovician Period. According to Sepkoski (1981b, 1995, 1997), after the dramatic Early Cambrian radiations there was a phase of apparent “stagnation” or quiescence that persisted through the Mid Cambrian to the Early Ordovician, with a comparatively lower overall diversity thoughout (figure 1.1B). Yet it remained an interval of high turnover—with high extinction rates limiting the overall buildup of diversity, including the well-defined successive pulses of mainly trilobite-based Mid to latest Cambrian “biomere” extinctions (Palmer 1979; Ludvigsen 1982; Zhuravlev 2000). The last of these more or less coincides with the base of the North American Ibexian (Ross et al. 1997), that is, three conodont zones below the base of the Ordovician (Cooper et al. 2001). A few extinction horizons have been identified within the Ordovician, but they all appear to be relatively minor pulses that have yet to be demonstrated as being truly global events. For example, Ji and Barnes (1993, 1996) documented an Early Ordovician (mid Tremadocian) conodont extinction event in Laurentia; Sepkoski (1992b, 1995, 1996) reported extinction horizons in the late Mid Ordovician (near the bottom and top of the Darriwilian, respectively); and Patzkowsky et al. (1997) identified a climatic event with associated Late Ordovician (mid Caradoc) brachiopod-dominated extinction in Laurentia. The last has wider, probably even global, importance, correlating with climatic and other changes—including the “late Keila” extinctions of chitinozoans, acritarchs, and ostracodes—in Baltoscandia (Kaljo et al. 1996; Ainsaar et al. 1999).

Biotas This book was compiled by a large team of Ordovician specialists from around the world, under the aegis of the joint UNESCO and International Union of Geological Sciences (IUGS) geosciences initiative—the International Geological Correlation Programme (IGCP)—which supported IGCP project no. 410. The aim in establishing this first globally oriented, internationally sponsored IGCP project was to have the “Great Ordovician Biodiversification Event” comprehensively evaluated in a collective effort. When IGCP 410 became formally established in

1997, data collection and analysis of biotas were coordinated mainly on a regional basis, but in 1998 a separate study program was developed with a more constrained global approach on the major fossil groups. This became an evaluation of how each taxonomic group diversified through Ordovician time, with assessments of patterns of diversity change, and rates of origination and extinction, based on the assembly of genus- and species-level taxonomic diversity data of the various fossil groups. Leading specialists and their colleagues were invited to participate in the three-year work program and to attend a major meeting organized by Mary Droser in June 2001 at the University of California in Riverside to discuss the results of the compilations. The book derives in part from these contributions, with a number of additional studies from other specialists added to the project after the Riverside meeting to widen the coverage to include nearly all fossil groups. The leader of each fossil group chose his or her own team of co-workers to tackle the compilation of data and assembly of the manuscript for the particular chapter. The groups varied in size from large (with a coordinating author and many coauthors) to small, single-author presentations. We encouraged each team to establish its diversity surveys so that they highlighted patterns of diversity change, originations, and extinctions, where possible down to species level at least for the pelagic groups, and to genus, and where possible to species level for benthic groups. Most authors assembled their primary taxonomic data using their own databases (spreadsheets or census lists). The pelagic groups (e.g., graptolites, chitinozoans, radiolarians) and some benthic groups (e.g., bryozoans, sponges, stromatoporoids, echinoderms) are presented as species-level, or combined species- and genus-level, diversity surveys. Larger benthic groups (e.g., trilobites, brachiopods, gastropods, bivalves, nautiloids) have been surveyed mainly at the genus level at this stage. Although we attempted to provide the widest possible coverage of Ordovician biotas, it was inevitable that some unevenness in the levels of documentation and analysis would occur from chapter to chapter. For a variety of reasons, a number of groups are presented with rather incomplete global diversity analyses. This may be because (1) only part of the group’s Ordovician record has been treated; (2) reliable data of well-preserved and diagnostic material are

3

4

 .  available only from one or two regions in the world, so the survey does not have a global focus; (3) genusand species-level assignments of a group are unreliable or at least need further revision before a worldwide analysis can be attempted; or (4) present-day expertise on a particular group is lacking. Consequently, for a few groups, only preliminary statements could be included, and in some cases these assessments remain predominantly regionally based. Although we have attempted to include most Ordovician groups in this genus- and species-level coverage (table 1.1), there are some gaps in the documentation— for instance, the Mid to Late Ordovician conodont diversity record and a few small groups of comparatively limited Ordovician occurrence, for example, foraminiferans (Lipps 1992a, 1992b), hydroids (Stanley 1986; Foster et al. 1999), and hyolithelminths. In addition, a few groups of microorganisms—such as two unicellular planktic green algal prasinophyte groups (leiospheres, tasmanitids) and the prolific, organic-walled, benthic cyanobacterium Gloeocapsomorpha prisca, a microorganism of considerable economic interest because it forms significant matlike accumulations of late Mid to Late Ordovician “kukersite” oil shales in the intracratonic basins of Estonia (Körts 1992), North America (Jacobson et al. 1988), and Australia (Foster et al. 1990)—have not been included in the survey. An estimate of total numbers of Ordovician genera and species for each fossil group treated in the book (table 1.1) is provided by authors for general guidance only. It is not intended as a comprehensive listing of totals for all Ordovician groups. The list comprises about 4,605 genera (excluding the trace fossils). The total number of animal genera (less the “plants”—algae, acritarchs, miospores) is 4,254 genera. Earlier, Sepkoski (1995) employed a database of 4,367 animal genera (i.e., 12 percent of all known animal genera in the Phanerozoic) to prepare his outline of Ordovician diversity history. Only the few small groups mentioned earlier, the late Mid to Late Ordovician conodont record and about 70 genera of nautiloids from China and Russia, have been excluded from our present survey. With recent publication of Sepkoski’s (2002) comprehensive listing of all known fossil marine animal genera, it is now possible to make more-up-to-date

TABLE 1.1. A Preliminary Listing of Genus- and Species-Level Totals for the Ordovician Fossil Groups Fossil Groups

No. of Genera

No. of Species

Radiolarians Sponges Stromatoporoids Conulariids Corals Bryozoans Inarticulated brachiopods Articulated brachiopods Polyplacophoran mollusks Symmetrical univalves Gastropods Bivalves Rostroconchs Nautiloid cephalopods Hyoliths Cornulitids Coleoloids Sphenothallids Byroniids Tentaculitids Scolecodonts (polychaete jaws) Chaetognaths Palaeoscolecidans Machaeridians Trilobites Eurypterids Ostracodes Phyllocarid crustaceans Echinoderms Chitinozoans Graptolites Conodonts (Early to Mid Ordovician) Vertebrates Receptaculitids

24 135 29 12 128 169 190 540 15 75 140 136 21 305 (375) 32 3 3 1 1 1 50 4 8 7 842 6 559 9,+4? 431 35 192 106 35 9

98 280 155 51 (97) n.d. 1,120 n.d. n.d. 31 c. 700 c. 2,500 530–680 107 n.d. n.d. n.d. n.d. n.d. n.d. n.d. 500–1,000 7 21 65 n.d. 9 +2,600 22,+9? 1,216 397 800 430 37 48

Algae (including cyclocrinitids) Acritarchs Miospores: cryptospores : trilete spores Land plants (no megafossil record)

88 250 12 1 n.d.

+164 c. 1,300 35 1 n.d.

Trace fossils (ichnogenera)

49

n.d.

Source: Compiled from data supplied by individual authors. Note: This compilaton of animals, plants, and trace fossils is provided for general guidance only because some genus-level and especially the species-level taxonomy of fossil groups remains in a state of flux. For many groups a large backlog of taxonomic revision work still needs to be done. Bivalve and scolecodont species data are shown as ranges, namely, between 530 and 680 and between 500 and 1,000 species, respectively. For entries of nautiloid genera and conulariid species, two numbers are included (one in parentheses): the lower number is the actual count (a mainly regional summation), and the higher number in parentheses is the global estimate. For the phyllocarids, the valid genera and species numbers are shown, plus the numbers of doubtful taxa (with a question mark). For trace fossils, only numbers of form genera (ichnogenera) could be supplied. Abbreviations: c. = about; n.d. = data not determined (or not supplied).

comparisons between the two lists of Ordovician animal generic diversity records (table 1.2). However, it is difficult to compare the two lists in detail because they represent completely different types of fossil data assemblies. The present study was not intended to be comprehensive but rather to involve as many available Ordovician specialists as possible in a stocktaking of their fossil group. In some cases these workers have

Introduction TABLE 1.2. Comparative Generic Lists of Ordovician Animal Diversity Data Animal (and Animal-like Protist) Fossil Groups

Genera: This Study

Genera: Sepkoski (2002) List

Protozoans: radiolarians Foraminiferans Spirotrichs (ciliates) Sponges Stromatoporoidsa Problematic receptaculitidsb Cnidaria–Medusae/Melanosclerites Conulariids Hydrozoans Corals Bryozoans Inarticulated brachiopods Articulated brachiopods Polyplacophoran mollusks Symmetrical univalvesc Gastropodsc Bivalves Rostroconchs Scaphopods Nautiloid cephalopods Tubelike taxa: hyoliths Cornulitids and coleoloids Sphenothallidsd Byroniidse Tentaculitids Hyolithelminthes Incertae sedis Scolecodonts (polychaete jaws ) Other Polychaete worms Palaeoscolecidan worms Myzostomian “worms” Pentastome “worms” Problematic machaeridians Arthropods: eurypterids Other merostomes Mimetasterids Trilobites Phyllocarid crustaceans Ostracodes Class incertae sedis Echinoderms Chitinozoans Graptolites-Graptoloids Other hemichordates Conodonts Chaetognaths (arrow worms) Vertebrates

24 135 29 9 12 128 169 190 540 15 75 140 136 21 375 32 6 1 1 1 50 8 7 6 842 13 559 431 35 192 106 (partim) 4 35

25 19 1 126 25 3 16 13 157 229 154 601 18 34 139 146 20 1 436 28 6 1 6 13 36 14 7 1 1 9 5 6 1 833 11 462 6 407 33 119 78 133 1 11

4,327

4,391

Total

Source: Compiled from data supplied by individual authors (this study) and from Sepkoski (2002). a Stromatoporoids are here listed (in this study and in Sepkoski’s list) as separate from other sponges. b The problematic receptaculitids is the only group omitted from Sepkoski’s list. They are included in this study as a possible metazoan group. c The bellerophontids have been included with the symmetrical univalves of this study, whereas they are included in Sepkoski’s list with the gastropods. d The genus (Sphenothallus) is included in Sepkoski’s list as a conulariid. e The genus (Byronia) is included in Sepkoski’s lists as a hyolithelminth.

adopted revised taxonomic categories from those listed in Sepkoski’s Compendium of Fossil Marine Animal Genera (2002). The Compendium represented a mammoth compilation of Phanerozoic animal genera, assembled from the data coverage of animal groups in Moore et al., Treatise on Invertebrate Paleontology (1953–1992), and many hundreds of other literature sources published up to 1998. Nevertheless, there is a remarkable degree of similarity between the two lists. The comprehensive Sepkoski (2002) compilation also provides a means to appreciate the full range of Ordovician taxonomic groups, including those biodiversity records that have not been documented here. Only a comparatively few groups have good-quality assignments of species data (table 1.1). These include mainly pelagic groups such as the graptolites, conodonts, chitinozoans, and radiolarians, but correct species assignments have also been established for a few benthic groups (e.g., stromatoporoids, conulariids, bryozoans, rostroconchs, eurypterids, echinoderms, and receptaculitids). Wilson (1992) has an estimate of 1.4 million organisms (plants, animals, and microorganisms) currently known to be living on earth, though millions of insects, microbes, and other organisms remain undescribed across habitats from rain forests to the ocean deeps (Wilson 1992; Thorne-Miller 1999). The problems of estimating fossil species numbers is a more daunting task. Paul (1998) reviewed the available approaches, namely, to establish what proportion of living animal species is likely to be preserved in modern settings, as well as to establish the relationships between soft-bodied and skeletonized organisms in fossil Lagerstättern such as the Middle Cambrian Burgess Shale (Conway Morris 1986). These two approaches produce rather similar results, suggesting to Paul that only about 10 percent of all Phanerozoic species are likely to have been preserved. In assessing the Ordovician global record, we currently have only reasonably complete numbers for the genera—some 4,600 known Ordovician genera. From the generic and specific data in the 26 listed fossil groups (table 1.1), there are on average about five species for each genus. Multiplying this value by the generic total gives the very approximate total of 23,000 Ordovician skeletonized species. If this estimate represents only 10 percent of species likely to be preserved (Paul 1998), then

5

6

 .  the overall total (including the soft-bodied organisms) may be of the order of 200,000 to 250,000 Ordovician species.

Ordovician Time and Time Slices A specific aim of IGCP 410 was to develop a more highly integrated, well-calibrated Ordovician timescale capable of providing a more reliable global and regional basis for correlating range data and establishing age relationships of the diverse biotas. It has long been recognized (Jaanusson 1960, 1979) that establishing biostratigraphic correlations using Ordovician biotas, especially on a global scale, is a difficult task because of the extensive biogeographic and ecologic differentiation of the faunas. Furthermore, the comparative lack of good reliable radiometric ages has limited the degree to which age relationships could be determined (Webby 1998). The priority to establish a stabilized global stratigraphic framework for the Ordovician System has been the main responsibility of the International Subcommission on Ordovician Stratigraphy (ISOS). It is an independent subcommission of the International Commission on Stratigraphy (ICS), and both ISOS and ICS exist under the umbrella of IUGS. Since the mid-1980s, ISOS has made good progress toward the establishment of a single set of globally based Series and Stage divisions, using the most highly resolved biologic indices (graptolite, conodont, and chitinozoan zones) coupled with available physical and chemical tools (table 1.3). Stratigraphers, especially those closely associated with global correlation work of international stratigraphic bodies like ICS and ISOS, have long been aware of the need to draw clear distinctions between “chronostratigraphic units” based on rock sequences and “geochronologic units” that reflect intervals of geologic time (Salvador 1994). The terminology is straightforward for the bulk of conventional hierarchical usages, such as “system,” “series,” and “stage” with their geochronologic counterparts “period,” “epoch,” and “age.” However, most “series” have tripartite subdivisions, with names derived from their position in the “system,” for example, “Lower Ordovician Series,” “Middle Ordovician Series,” and “Upper Ordovician Series.” The nomenclature “Early Ordovician Epoch” and “Late Ordovician Epoch” is also acceptable, but

the terminology for the intervening epoch has remained a problem because the same name “Middle Ordovician” has been applied to both series (for the rocks) and epoch (for time) usages. Only British stratigraphers (see Holland et al. 1978) have drawn a distinction in their regional work between “middle” (for rocks) and “mid-” (for time). According to S. C. Finney, who is the current chair of ISOS and second vice-chair of ICS, the terms “Middle Ordovician” and “Mid Ordovician” are now “used for the series and its correlative epoch, respectively” (Finney et al. 2003:351). Following such recent moves at the international level, we have adopted the formalized global names “Middle Ordovician Series” for the rock sequences and “Mid Ordovician Epoch” for time throughout this book. The categories of lower, middle, upper (for rocks) and early, mid, late (for time) have similarly been recognized for ratified, tripartite, global stage subdivisions and for informal, mainly lower-rank, usages, as depicted in table 1.3. Of critical importance, especially for biodiversity studies, was the level of resolution of the timescale (Erwin 1993; Conway Morris 1998a). Clearly it was essential to establish the most refined timescale with the shortest-term subdivisions for the widest applicability in studies of global biodiversity change. Consequently, two new, highly integrated and resolved correlation charts were prepared by a small group of specialists in IGCP 410, in cooperation with members of ISOS. These present the global and regional stratigraphic frameworks, main biostratigraphic (graptolite, conodont, and chitinozoan) schemes for establishing cross ties, key radiometric dates, and a set of 19 time-slice divisions by Webby et al. (see chapter 2: figures 2.1, 2.2). Sadler and Cooper’s (chapter 3: figure 3.1) new calibration of the Ordovician timescale based on computer-assisted constrained optimization (CONOP) procedures allowed the subdivisions in the correlation charts to be calibrated with respect to geologic time. CONOP calibrations involved processing a huge data set of first- and last-appearance events based on 1,100 or so Ordovician and Silurian graptolite species from nearly 200 measured sections. A “best fit” sequence of events was derived in the CONOP process to establish a relative timescale from the thousands of event levels. The scale became numerically calibrated when radiometric dates and ranges of key graptolites

Introduction TABLE 1.3. Summary of the Global Subdivisions of the Ordovician System/Period System/Period

Series/Epoch

Stage/Age

Substage/Subage

Time Slices (TS)

Ashgill

upper/latea middle/mid lower/early

6c 6b 6a

Caradoc

upper/late middle/mid lower/early

5d 5b–c 5a

Darriwilianb

upper/late middle/mid lower/early

4c 4b 4a

lower Middle/early Midc

N/A

3a 3b

upper Lower/late Earlyc

N/A

2c 2b 2a

Tremadocian

upper/late middle/mid lower/early

[443Ma] Upper/Late O R D O V I C I A N

Middle/Mid

Lower/Early [489Ma]

1d 1b–c 1a

Ratified global subdivisions are shown in bold and are capitalized; informal stage/age and substage/subage usages are not capitalized; estimated radiometric ages (Ma) for the base and top of the Ordovician are shown in square brackets; N/A = subdivisions not applicable. Note that a distinction is drawn between the “System,” or body of rocks, and the “Period” representing geologic time. The 19 time slices adopted for use in this book are shown in the right column. a The British Hirnantian stage is equivalent to the upper/late Ashgill interval. b The British Llanvirn is correlative with the middle/mid to upper/late parts of the Darriwilian Stage. c The British lower/early Arenig and the middle/mid Arenig are more or less correlative with the unnamed upper Lower/late Early Ordovician and lower Middle/early Mid Ordovician global stages/ages, respectively.

(and conodonts) were interpolated along a regression line. In a more recent development, Cooper and Sadler (2002) have shown that the CONOP application can be used directly, without the “time unit bias,” to establish a global diversity curve. Again taking the wellstudied and widely distributed graptolites, they have depicted a “running diversity curve” of 2,272 estimates of “interval free” graptolite species standing diversity through Ordovician and Silurian time. The CONOP calibrations (chapter 3) for this biodiversity project have added significantly to establishing the Ordovician timescale as an integrated, highresolution dating and correlation tool, to a point where it is now arguably one of the best-resolved intervals of time within the Paleozoic Era. This indicates, despite Jaanusson’s (1960, 1979) earlier misgivings, that the extreme levels of biogeographic and ecologic differentiation of the Ordovician biotas were not limiting factors. The 19 slices employed for the Ordovician Period (46 m.y. duration; table 1.2), with time-slice intervals of relatively equal units of time (intervals that each span between 1.6 and 3.0 m.y.), provide a most practical, best-resolved, and reliable basis for use in this present biodiversity survey. These standardized time slices, as emphasized by Cooper (1999c:441),

are “as fine as is practical” to allow precision in global studies of diversity change of the fossil groups. Two types of abbreviations are employed through this work to distinguish between time, as a duration or interval in millions of years (m.y.), and time as a specific age assignment including radiometric ages in millions of years (Ma). Authors of the chapters in this volume were provided with the correlation charts (chapter 2: figures 2.1, 2.2) so that time equivalences could be maintained in data compilations across the different groups. In addition, abbreviated left- and bottommargin versions of the key elements of these Ordovician stratigraphic charts were supplied so that they could be incorporated at the margins of diversity plots when compiling diversity data and printed with a consistent time frame for comparative purposes throughout the chapters of the book. The majority of authors have, in compiling their diversity data, successfully employed the standardized global time-slice scheme across a wide range of environmentally and provincially differentiated profiles.

Diversity Measures Another important aspect of this biodiversity study was to have some consistency in the kinds of measures

7

8

 .  used in the volume for summarizing the data on diversity patterns and turnover rates. A short chapter on diversity measures is contributed by Cooper (chapter 4). The guidelines distributed to contributors provided general recommendations for establishing the best estimate of the “true mean standing diversity,” irrespective of the length of the time unit and taxonomic rank (Cooper 1999c), and the means of determining origination and extinction rates, the best approach being to estimate the rates in million-year intervals, which, of course, is dependent on the reliability of the calibration of the timescale (see earlier in this chapter). The measures have been adopted by about half the contributors, giving overall a relatively uniform set of presentations without limiting flexibility for those authors who wished to use other measures to emphasize particular features in their databases. Some additional measures that were circulated in a final update of the guidelines could not be adopted by some authors because they had already compiled their data and submitted their manuscripts. Cooper’s standardized symbols (d, o, e, and subscript i, for diversity, originations, extinctions, and interval in m.y.) have been used widely throughout the book. On the other hand, no symbols were recommended to denote the taxonomic level of the compilation (genus or species), as it was thought preferable to explain the hierarchical level of the compilation in the text or in an appropriate caption. Cooper’s (chapter 4) contribution is important in identifying “normalized diversity” (dnorm ) as consistently providing the best estimate of “true mean standing diversity”; hence it is the most appropriate measure for determining diversity change through geologic time. With the highly resolved and calibrated global and regional time frame, the contributors of the various fossil groups were able to plot their data (genera and/ or species) accurately as global and/or regional stratigraphic ranges and then, using the diversity measures provided, derive the patterns of diversity change and turnover rates from their plotted range-chart data. The diversity data have been, in some cases, plotted in two sets of diversity curves, one based on actual “sampled” diversity through time slices and the other as “range-through” diversity, whereby a taxon is shown as having a continuous range, though absent from a time slice (possibly owing to sampling failure or bio-

facies shift) between occurences in the immediately preceding and succeeding time slices (for examples, see Regional Patterns in Australia and New Zealand, the Anglo-Welsh Sector of Avalonia and in South China, chapter 24). ■

Organization

The present book derives from the contributions of 96 Ordovician paleontologists, stratigraphers, geochemists, and other geologists from 17 countries throughout the world. The book comprises this introduction and thirty-four other chapters, which have been divided into four main sections: part I: Scaling of Ordovician Time and Measures for Assessing Biodiversity Change; part II: Conspectus of the Ordovician World; part III: Taxonomic Groups; and part IV: Aspects of the Ordovician Radiation. The three chapters of part I deal with the stratigraphic framework, time slices, and calibration of the Ordovician timescale (chapters 2 and 3) and the measures of diversity (chapter 4) that formed the basis of this biodiversity study. Outlines for these chapters have been presented earlier. The remainder of this introductory chapter comprises (1) a brief history of earlier work on Ordovician biodiversity topics and (2) an overview of all the remaining chapters, divided into two sections, for part II (chapters 5–10) under the subheading “Ordovician World in Brief ” and for parts III and IV (chapters 11–35) under the subheading “Synopses of Fossil Groups.” Part II (Conspectus of the Ordovician World; chapters 5–10) includes a broad overview of a number of significant physical and chemical features of the Ordovician world, presented in order to (1) illustrate what marked contrasts exist between the Ordovician and the present-day climates and in the dispositions of continents and oceans and (2) provide background to some of the features that may have been influential in shaping this greatest diversification of marine life. Part III (Taxonomic Groups; chapters 11–33) comprises a nearly complete coverage of taxonomic groups. For reasons discussed earlier in this introduction, the chapters vary considerably, from comprehensive, integrated surveys of global and regional genus- and/or species-level diversity data to far more preliminary compilations, focusing mainly on a few aspects of

Introduction regional genus- or species-level biodiversity. The two concluding chapters of part IV (Aspects of the Ordovician Radiation; chapters 34 and 35) deal with the ichnofossil record and a global synthesis of the Ordovician Radiation. ■

Ordovician Biodiversity Perspectives: Earlier Work Evolutionary Faunas (EFs) and Floras: Some Global Considerations Faunal Patterns

Sepkoski’s (1978, 1979, 1981a, 1984, 1991a, 1997) numerous contributions on the patterns of the Phanerozoic global marine animal diversity change have given much attention to the development of the three “Great Evolutionary Faunas” (Cambrian, Paleozoic, and Modern EFs) and the pivotal role played by the Ordovician Radiation. The representatives of these EFs were recognized by Sepkoski as somewhat “fuzzy” sets of unrelated higher taxa with similar histories of diversification through extended intervals of time. At each level of taxonomic hierarchy (order, family, or genus), there were marked differences in the relative intensities of the Cambrian and Ordovician radiations. In comparison with the Cambrian increases, approximately twice as much ordinal biodiversity was added to the Ordovician marine system, some three times more familial diversity (figure 1.1A), and nearly four times more genus-level diversity (Sepkoski 1978, 1988; Sepkoski and Sheehan 1983). Each successive EF exhibited a greater diversity and ecologic complexity (Sheehan 2001a). The Cambrian EF was an explosive evolutionary event involving appearances of the first wellskeletonized metazoans—groups such as trilobites, inarticulated brachiopods, hyoliths, “monoplacophorans,” and eocrinoids. The fauna expanded exponentially during the Early Cambrian, an interval of 20 m.y. (Sepkoski 1979, 1997: figure 1; Zhuravlev 2000); leveled off during the Mid to Late Cambrian, at a peak diversity of about 100 families; and then declined gradually through the Ordovician as the Paleozoic EF expanded. The Cambrian EF was reduced still further to about 30 families by the short-lived, glacially induced end Ordovician mass extinction. The fauna included appearances of a larger number

of higher-level categories—most of the skeletonized metazoan phyla and nearly two-thirds of the classes (Sepkoski 1981a). On the other hand, lower-level categories were sometimes depauperate—“rather plain, even grubby” species, according to Valentine (1973: 451). The communities were mainly benthic, rather generalized, and intergrading, composed of surface deposit feeders, grazers, carnivores, and suspension feeders with low epifaunal or infaunal tiering. Overall the fauna exploited little of the potential ecospace (Bambach 1983). This contrasts markedly with the patterns of global diversity change during the Ordovician Period. The main contributors in this new wave of major diversification were elements of Sepkoski’s (1979, 1984) Paleozoic EF. The fauna included the articulated brachiopods, cephalopods, crinoids, ostracodes, stenolaemate bryozoans, and corals. Components of the Paleozoic EF diversified very slowly in the Cambrian, with only slight increase during the Late Cambrian. In the Ordovician an exponential rise of diversity occurred, which was at a more rapid rate than at any other time during the Phanerozoic (Sepkoski 1995). The communities of attached epifaunal suspension feeders, deep burrowers, and carnivores greatly expanded (Bambach 1983; Droser and Sheehan 1997a), as did more specialized reef and hardground communities. Compared with those of the Cambrian EF, the taxa of the Paleozoic EF became more specialized, with the development of narrower ecologic requirements, use of resources, and competitive abilities. Suspension feeders of all types began to appear, along with the scavengers and carnivores that preyed on them from the earliest Ordovician (Signor and Vermeij 1994). More than 350 new “Paleozoic” families were added in what Sepkoski (1981b:204) called the “largest turnover in composition of marine faunas seen in the history of the oceans” (figure 1.1A). The familial diversity curve, then, exhibits a sudden flattening after its steep upward climb, and this is interpreted to mean that the ecospace had finally been filled to capacity, to its equilibrium level, as shown by the nearly horizontal, Paleozoic-wide diversity plateau (Sheehan 2001a). However, two extinction/rebound perturbations disrupted the continuity of this diversity plateau; the first was the end Ordovician mass extinction, and the second was represented by the pulses of Late Devonian extinction. Despite these, the Paleozoic EF remained

9

10

 .  dominant for nearly 240 m.y. until it was far more severely disrupted in the end Permian mass extinction (Erwin 1993). Zhuravlev (2000) noted that, with the exception of the bryozoans, all the phyla that participated actively in the major Ordovician diversifications were present in the Cambrian (e.g., precursors to chordates, conodonts and graptolites, the trilobites, brachiopods, sponges, echinoderms, and the mollusk groups, such as rostroconchs, cephalopods, gastropods, and polyplacophorans). Most of these groups (apart from trilobites and inarticulated brachiopods and, to a lesser extent, sponges and echinoderms) remained unimportant in Cambrian communities, but they gradually emerged from the shadows as opportunities arose following the repeated extinctions of late Early, Mid, and Late Cambrian time (Zhuravlev 2000). By the latest Cambrian (Sunwaptan) they had already taken over, in the sense that the last diversity peak was produced by them. Sepkoski’s (1979, 1984) Modern EF had a comparatively limited development in the Ordovician, though its two dominant mollusk members (gastropods and bivalves) did expand markedly during Ordovician time (Sepkoski and Sheehan 1983; Sepkoski 1991a). Crustaceans, gymnolaemate bryozoans, foraminiferans, echinoids, fishes, reptiles, and mammals were other typical Modern EF components. The fauna exhibited a greater variety of predators and more highly complex infaunal communities (Bambach 1983, 1985; Thayer 1983). Such changes allowed ecospace to be divided more finely and still higher diversities to be achieved. The changes through the Ordovician Period mainly involved a very slow and steady increase in importance of members of the Modern EF (mainly bivalves and gastropods) relative to the overwhelmingly dominant members of the Paleozoic EF. This pattern continued through the Paleozoic (figure 1.1A), until the end Permian mass extinction, when differential survival of the faunal components occurred—the Paleozoic EF losing 79 percent of its familial diversity, compared with the Modern EF’s 27 percent loss (Sepkoski 1984). The Modern EF became the dominant fauna, and the expansion continues to the present with no sign of the appearance of a diversity plateau (Sepkoski 1997).

Overview of the Ordovician Diversity Record Sepkoski (1995) presented a very important summary of the diversification history of Ordovician biotas that focused not only on all the marine animal genera (see figure 1.1B) but also on typical representatives of individual fossil groups, with separate diversity curves for each group. Figure 1.1B is basically a reproduction of Sepkoski’s original (1995) figure 1— that is, it includes his primary data compiled in Ordovician generic-level diversity curves of marine animal genera against 12 Ordovician “time units” derived from British series subdivisions, with a scale bar added at the bottom of his diagram, for comparative purposes, to illustrate relevant new global Lower, Middle, and Upper Ordovician series, radiometric dates, and time-slice data as used throughout this volume. Sepkoski (1995) established his 12 “time unit” subdivisions by dividing each British series into two, except for the “Llandeilo,” which he left undivided, and the Caradoc, which he divided into three. In addition, his lower Arenig and upper Ashgill subdivisions were allocated longer and shorter durations, respectively, in comparison with the other time units. Similar stratigraphic subdivisions have been used in Sepkoski’s (1992a, 2002) family- and generic-level compendia. Stratigraphic relationships between Sepkoski’s subdivisions and the global series and time-slice (TS ) subdivisions used in this volume are shown in figure 1.1B, in particular, the base of the Middle Ordovician with a position toward the middle Arenig, and the base of the Upper Ordovician with a level close to the middle “Llandeilo.” Sepkoski’s (1995) figure 1 (also reproduced here in figure 1.1B) for all the marine animal genera illustrates that the global diversity levels during the Early Ordovician were maintained at levels broadly similar to those of the preceding late Mid to Late Cambrian. Both of these Cambrian and Early Ordovician intervals were interpreted by Sepkoski (1995) as intervals of high turnover (i.e., including high rates of extinction, as discussed earlier) with several “Paleozoic” EF groups expanding slowly at the expense of Cambrian EF members through Late Cambrian to Early Ordovician time. This pattern contrasts with representations of the Ordovician part of the diversity trajectory

Introduction shown in Sepkoski’s Phanerozoic-wide compilations. In both generic-level plots (see Sepkoski 1997: figure 1-1, 1998: figure 2) and familial-level plots (see Sepkoski 1979: figure 7, 1984: figure 1, 1997: figure 1-2), the dramatic, exponential rise of diversity is shown commencing near the beginning of the Ordovician and continuing more or less unabated—the slope of the trajectory remaining steeply inclined upward—until leveling off in the late Caradoc–early Ashgill (see Sepkoski 1984: figure 1, reproduced here as figure 1.1A). The first major diversification pulse of “Paleozoic” groups commenced early in the Mid Ordovician (mid Arenig—TS.3a–b) and continued to an initial early Darriwilian (late Arenig—TS.4a) peak; then there was a short-lived mid Darriwilian (or Llanvirn— TS.4b) lag that Sepkoski linked to the North American “Middle Ordovician” Sauk-Tippecanoe sequence boundary (an interval, or intervals, of lowered sea level highstands; see Ross and Ross 1995; Golonka and Kiessling 2002). The second major diversification started during the late Darriwilian (formerly latest Llanvirn–early “Llandeilo”—TS.4c) and reached its culmination in a mid Caradoc (TS.5b–c) maximum. A limited decline followed in the late Caradoc (TS. 5d), before rapid increase resumed to the sharp, third and highest diversity peak during the early to mid Ashgill (TS.6a–b). The late Ashgill mass extinctions then caused the dramatic decline of the diversity, virtually back to levels previously attained in the early Mid Ordovician, near the start of the first of the three great diversification pulses of the Ordovician Radiation, some 25 m.y. earlier. The second feature of Sepkoski’s (1995) documentation is the presentation of individual plots for some of the major taxonomic groups, the Cambrian EF represented by two groups (trilobites and inarticulated brachiopods), the Paleozoic EF by six groups, (articulated brachiopods, echinoderms, corals, bryozoans, nautiloid cephalopods, and rostroconch mollusks), the Modern EF by two groups (gastropods and bivalves), and two other mainly pelagic groups (graptolites and conodonts) that Sepkoski assigned to the Paleozoic EF. Some features of these patterns are outlined here, but detailed comparisons of his curves with counterparts documented in chap-

ters of this book are not attempted because of the differing fossil data assemblies, taxonomic categories, and methods of presentation employed by many authors. For the Early Ordovician record of more than 550 genera, the trilobite and inarticulated brachiopod components of the Cambrian EF make up about 49 percent of the total; the articulated brachiopod, nautiloid, rostroconch, bryozoan, graptolite*, and conodont* components of the Paleozoic fauna constitute about 42 percent of the total (14 percent of them pelagics—the groups with asterisks); and the gastropod and bivalve components of the Modern fauna make up about 8 percent of the total. Of these various groups, the rostroconchs have their main diversification through the Late Cambrian to Tremadocian interval, so this is one small “Paleozoic” group that started to diversify in the Late Cambrian, reaching its maximum in the Early Ordovician. Among other “Paleozoic” groups, at least one order (Ellesmerocerida) of nautiloids diversified in successive pulses across the Late Cambrian to Tremadocian boundary, and the “pelagic” graptolites and conodonts also radiated markedly during the Early Ordovician. This latter fact suggested to Sepkoski that, in the early history of the Paleozoic EF, some decoupling of the pelagic realm from the benthic realm may have occurred. The marine generic diversity doubled to about 1,200 genera during the Mid Ordovician, and there were significant changes to the proportions of the three EFs. The Cambrian EF (trilobites and inarticulated brachiopods) was now represented by only about 30 percent of the total, whereas the Paleozoic EF (articulated brachiopods, echinoderms, corals, bryozoans, cephalopods, rostroconchs, graptolites*, and conodonts*) had expanded to about 62 percent of the total (nearly 12 percent of them pelagics). The Modern EF (gastropods and bivalves) remained a minor component, as in the Early Ordovician, constituting close to 8 percent of the total. A significant feature of the Paleozoic EF is that benthic elements (especially articulated brachiopods, cephalopods, echinoderms, and bryozoans) expanded more rapidly than the pelagic components through the Mid Ordovician. On the other hand, of all the groups, only the graptolites

11

12

 .  and cephalopods are shown attaining their maximum generic diversities during the Mid Ordovician. Frey et al. (chapter 21), however, show the cephalopods as reaching a relatively slightly higher maximum diversity in the Late Ordovician. The Late Ordovician record comprises two greatest diversity increases, the first as shown in the broad, gently arching diversity curve of increase to a mid Caradoc maximum of about 1,600 genera, and the second in a higher, sharper, mid Ashgill peak of nearly 1,800 genera (figure 1.1B). It seems that all the groups surveyed by Sepkoski (1995), except apparently the rostroconchs and graptolites, attained their highest Ordovician generic diversities during this epoch. The Cambrian EF (trilobites and inarticulated brachiopods) had now declined further to about 23 percent of the total, while the Paleozoic EF (articulates, echinoderms, corals, bryozoans, cephalopods, rostroconchs, graptolites*, and conodonts*) had expanded further to about 64 percent of the total (of which nearly 8 percent were pelagics). In addition, the Modern EF (gastropods and bivalves) had now increased slightly in importance, to 12 percent of the total. All groups exhibit a sharp major decline associated with the end Ordovician extinctions. The Late Ordovician diversity peaks of individual groups show some interesting patterns, with all the maxima associated with the early Caradoc (TS.5a), mid Caradoc (TS.5b–c), and/or the mid Ashgill (TS.6b) intervals. For the Cambrian EF groups, the trilobites have a diversity peak in the early Caradoc, while the inarticulated brachiopods have a mid Caradoc peak, though both groups exhibit a small, secondary mid Ashgill peak. Three of the Paleozoic EF groups (articulates, echinoderms, and bryozoans) show rather similar-sized mid Caradoc and mid Ashgill diversity peaks; the conodonts also show these same peaks, though they are less well defined. The nautiloids and graptolites, on the other hand, have their two prominent diversity peaks spaced farther apart, the first in the early Caradoc and the second in the mid Ashgill (TS.6b). The rostroconchs have only one minor early Caradoc peak, while the corals show only one prominent, major mid Ashgill peak. Of the components of the Modern EF, the bivalves show all three diversity peaks (early and mid Caradoc and mid Ashgill), and the gastropods show a steadily rising diversity curve to a single mid Ashgill peak.

Floral Patterns The Ordovician floral groups have received comparatively little attention. The calcified benthic marine Ordovician “algae” were originally grouped by Chuvashov and Riding (1984) in their “Ordovician Flora.” This was differentiated as one of the three major Paleozoic “evolutionary” floras based on characteristic Cambrian, Ordovician, and Carboniferous assemblages associated with reefs, stromatolites, oncoids, or “debris” in the shallow marine carbonates. The Cambrian flora was dominated by cyanophytes (blue-green “algae”) that appeared during a brief span (5 m.y.) of the earliest Cambrian. The “blue greens” are now commonly referred to cyanobacteria because they are prokaryotic and therefore related to bacteria (Riding 1991) but differ in having chlorophyll a and in developing a typical thallus without roots, stems, or leaves (for further discussion, see chapter 31). Riding (2000) outlined the cyanophyte radiation of the Cambrian flora, from its dramatic appearance in the earliest Cambrian through its progressive decline during the Mid to Late Cambrian into the Early Ordovician, when it finally more or less disappeared. The Ordovician flora—represented by mineralized (calcified) thallophytes—was dominated by dasyclads, codiaceans/udotaceans, and solenoporans (green and red algae). Chuvashov and Riding (1984) indicated that this flora appeared through the first two-thirds of the Ordovician Period. However, the major radiations of these groups took place later, mainly through Darriwilian to Caradoc time (see chapter 31). The great expansions of metazoans and calcified algae (“Paleozoic” fauna and Ordovician flora) had a profound impact on the “Cambrian” cyanophytes. Their decline is well documented in the reef habitat (Rowland and Shapiro 2002: figure 8; Webby 2002), but it is not yet known whether the disappearances relate to real extinctions or merely represent the loss of a preserved record when cyanophyte calcification processes ceased in unsuitable environmental conditions, possibly related to temperature decline (Riding 1992). The other floral groups are represented by nonmineralized, microscopic thallophytes, the acritarchs, and the dispersed spores of the first bryophytelike land plants. The thallophytes are unicellular green algae (e.g., prasinophytes) and form a small component of the organic-walled phytoplankton in the Ordovician

Introduction oceans. The acritarchs represent a major, organicwalled phytoplankton component of the oceans, but their identification as cysts of unicellular thallophytes remains uncertain (chapter 32). The cryptospores are predominantly dispersed spores and probably derived from small nonvascular bryophytelike plants (chapter 33). The acritarchs had an important early record in Proterozoic-Cambrian oceans—in the Early Cambrian up to 100 species (Vidal and Moczydlowska-Vidal 1997)—and then became significantly more diversified and abundant during the Ordovician (some 250 genera and about 1,300 species; table 1.1). Servais et al. (chapter 32) indicate that well-defined warmer water and cooler “provincial” assemblages were developed during the Early to Mid Ordovician. Based on British species data, there was apparently a progressive rise through the Early Ordovician to a Darriwilian diversity high, then continuous decline during the Late Ordovician. This contrasts with the Late Ordovician record in Baltoscandia, where Kaljo et al. (1996) have shown fluctuating, relatively high levels of diversity through the same interval, including localized peaks of diversity—two during the mid Caradoc and another in the early Ashgill. These somewhat divergent regional Mid to Late Ordovician acritarch pelagic diversity results do not closely match the record of mainly Darriwilian to Caradoc radiation events for the benthic calcified algae. Two separate diversification phases seem to have been responsible for the earliest land-based floras. The first is indicated by the appearances of moderately abundant and cosmopolitan cryptospore records from the Darriwilian onward, and the second is suggested by the record of trilete spores in the latest Ordovician, which possibly signals the emergence of the earliest vascular plants (chapter 33). From this timing, only the first of these events can be related to the main radiation events for the benthic, calcified algae. However, it seems more likely that all these different Ordovician floral radiation events occurred completely independently of one another in their separate benthic, pelagic, and terrestrial realms.

Evolutionary Faunas at the Ecologic Level North American and Other Continental Platforms Sepkoski (1981b, 1991a), Sepkoski and Sheehan (1983), and Sepkoski and Miller (1985) highlighted

the close relationships that exist between global diversity change within the successive “Evolutionary Faunas” (EFs) and localized to regional communitybased diversity change. The sequential diversifications of the three EFs were recognized in the onshoreoffshore patterns of expansion of the three basic trilobite-, brachiopod-, and mollusk-dominated community types. These were seen to be environmentally controlled and governed by successive onshore originations and offshore expansion, with replacement, or displacement, through time. Sepkoski and Sheehan (1983), Sepkoski and Miller (1985), and Sepkoski (1991a) employed about 500 Paleozoic (100 of them Ordovician) level-bottom “communities” across the North American platform for their compilations. They first converted the faunal lists to ordinal counts of generic (or species) numbers for each community and then used cluster and factor analyses to analyze the assembled data, with assembly of the results in a series of time-environmental “maps.” These “maps” depict the major changes through Ordovician time: (1) the more complexly structured Paleozoic benthic communities displacing preexisting, less-structured Cambrian communities to the outer shelf and slope; and (2) the Modern mollusk-rich communities originating onshore and, in turn, displacing the Paleozoicdominated communities to more offshore sites. Consequently, the Ordovician Radiation, with its significant components of the Paleozoic and Modern EFs in the brachiopod-rich and mollusk-rich communities, exerted a profound impact on the ecologic structure, with the older (Cambrian-type), and to a lesser extent the Paleozoic-type, communities being displaced diachronously offshore (Sepkoski and Sheehan 1983; Sepkoski and Miller 1985; Sepkoski 1991a). These events triggered great benthic community restructurings in the low latitudes of the North American continental shelf to slope during the Ordovician, but we are still some way from settling whether they were part of an overall global restructuring across the full range of geotectonic settings, facies profiles, and paleolatitudes or strictly regional (North American) events. Only by establishing similarly intensive, wide-ranging, and rigorous programs of Ordovician biodiversity study in other major platform regions such as Baltoscandia and South China can we necessarily expect to determine convincingly whether global processes were ultimately responsible for triggering these restructurings.

13

14

 .  On the other hand, Miller (1997a) has shown that promising results are attained if the global patterns are dissected—that is, if diversification patterns are compared at different geographic and environmental scales. His comparative database survey of more than 6,570 genus-level Ordovician occurrences of major faunal representatives of each of Sepkoski’s (1981a) three EFs—trilobite-, brachiopod-, and mollusk-dominated—across six different continental blocks (Laurentia, Baltoscandia, North China, South China, Bohemia, and East Avalonia) revealed distinctive continent-to-continent differences in generic richness and compositional changes between the faunas through Ordovician time. Both raw data and rarefaction analyses were used, the latter to compensate for the uneven sample sizes—each bin equivalent to one very unequal British series subdivision of time. Further comparative analysis of the Laurentian and South Chinese data was provided by Miller and Mao (1998) outlining the generic diversity patterns: (1) generally higher in Laurentia than in South China; (2) levels rather static in the Early to Mid Ordovician, then becoming higher in the Caradoc for Laurentia; (3) levels in the Mid and Late Ordovician probably somewhat higher than for the Early Ordovician of South China; and (4) increases of benthic mollusks only in the Late Ordovician of Laurentia. The major contrast in the patterns of diversification of the two continental areas relates to the presence or absence of siliciclastic-rich environments. Carbonate sedimentation was predominant in South China, but there were influxes of terrigenous sediments across eastern and central Laurentia during the Late Ordovician, associated with orogenic activity (Taconic Orogeny), and this facilitated the radiation of benthic mollusks, especially the bivalves (chapter 20). Miller and Mao (1998) recognized that the orogenic activity had importance as an abiotic factor in this continental-scale Late Ordovician radiation of benthic mollusks. In a more recent survey, Connolly and Miller (2002) have sought to explain global origination and extinction patterns by using diversity- and productivity-dependent models to analyze their standard data set of trilobite, brachiopod, gastropod, and bivalve occurrences. The bivalves were recognized as the only group to show productivity-dependent origination, as a consequence of the increased orogenic activity in the Late Ordovi-

cian, but this overall had a rather limited impact on origination patterns. Another community-based approach for studying Ordovician biodiversity has involved the assessment of taxonomic richness using two main types of measurement: (1) “inventory” diversity, which is usually represented by alpha (within habitat) diversity, or the numbers of taxa per unit area, and basically records taxonomic richness, or packing, in the habitat; and (2) “differentiation” diversity, which is a measure of the dissimilarity (or similarity) between components of the inventory diversity. Following Sepkoski (1988), this comprises beta (between habitat) diversity, or a measure of variation in taxonomic composition between areas of alpha diversity, and gamma (between province) diversity, a measure of taxonomic differentiation between geographic regions (a reflection of provinciality or endemicity). Sepkoski (1988) took the 500 or so Paleozoic assemblages from the North American platform—much the same database as used in the studies of faunal change mentioned earlier in this chapter—though he used a more precisely defined onshore-offshore framework encompassing six environmental “zones” for this compilation. The survey focused on the patterns of generic alpha and beta diversity, particularly the contributions during the Ordovician radiations. Averages of alpha diversity were computed across each environmental “zone” and for each geologic period. Each period exhibited a trend of alpha diversity that increased offshore to a high about midshelf then declined farther offshore. Sepkoski was only able to demonstrate an overall generic alpha diversity increase of 70 percent from the Cambrian to later in the Paleozoic (i.e., of diversity increase that could be related to Ordovician radiations). This value falls far below the 300 percent global generic diversity increase derived from the compilation of the marine animal genera based on the Treatise (Moore et al. 1953–1992) and other sources (Sepkoski 1986, 2002). Sepkoski (1988) sought a number of explanations for this missing diversity. Was it incorporated in the beta diversity or derived from other sources? The trends of beta diversity between the Cambrian and Ordovician suggest some increased habitat specialization, as well as addition of soft-bottom communities and appearances of new community types (reef and

Introduction hardground communities, bryozoan thickets, and crinoid gardens), but still the amounts are insufficient to relate to the missing diversity. Although Sepkoski (1988) discounted gamma diversity as having a role, it is apparent that the significant levels of provinciality that have been recognized in many groups of organisms through Ordovician time were overlooked (Jaanusson 1979; Patzkowsky 1995c; Webby et al. 2000). Moreover, Sepkoski’s assessments of alpha and beta diversity were based on regional North American community studies rather than global data. Miller (1997c) and Miller and Mao (1998) also noted that Sepkoski (1988) did not measure patterns throughout the Ordovician. Miller and Mao’s (1998) comparative analyses of data show beta diversity in South China declining markedly through late Mid to Late Ordovician time, whereas in Laurentia the beta diversity exhibits only a slight overall decline. The alpha diversity patterns also change through time, with both continental platform areas exhibiting modest increases through the latest Darriwilian to Caradoc interval. The global summation of community-scale diversification has yet to be determined because it requires assessments across all continental and oceanic terranes. Other Geotectonic Settings and the Oceanic Sampling Biases Caused by Subduction Ordovician community-based diversity change has been recognized in other geotectonic settings (especially the small continental blocks and terranes with oceanic and island-arc remnants), for example: (1) in the smaller, faunally distinctive, high-latitude periGondwanan terrane of Perunica–Bohemia (see Kraft and Fatka 1999); (2) in the low-latitude microcontinental block of the Argentine Precordillera (Waisfeld et al. 1999); (3) in the intraoceanic to marginal island sites of Newfoundland in the Iapetus Ocean (Harper and Mac Niocaill 2002); (4) in the Kazakhstanian terranes (Popov et al. 1997); (5) and in the periGondwanan Macquarie terrane with associated volcanic islands in eastern Australia (Webby 1992c; Percival and Webby 1996, 1999). All these geotectonically distinctive sites need to be well documented, and ecologically, and in a few cases provincially, distinct diversity profiles need to be fully developed.

Depending on their size, age, isolation, topography, and climate, islands and archipelagoes are potential sites for evolution and dispersal and are most important contributors to the biodiversity record (MacArthur and Wilson 1967; Soja 1992). Yet the patterns of ancient marine island biotas have been largely ignored. The greater part of the pre-Mesozoic oceanic crustal record (with its sedimentary and volcanic island records) has been destroyed by subduction (Menard 1986). Much of the preserved Ordovician record of the great Panthalassic Ocean (figure 5.4) that presumably straddled the Northern Hemisphere, for example, seems to have been lost. Nevertheless, some important oceanic and island-arc remnants have been incorporated into the terranes of Paleozoic fold belts. The oceanic deposits that survived the various accretionary processes are commonly metamorphosed and/ or structurally disrupted, but some of these “windows” may exhibit a reasonably well-preserved fossil record of oceanic (pelagic and deep benthic) life and shallower benthic faunas from the fringes of volcanic island sites. The importance of Ordovician islands as sites for evolution and dispersal was first emphasized by Neuman (1972). He noted the islands in the Iapetus Ocean as providing (1) a wide range of shallow-water habitats, (2) pathways for migration, (3) centers for evolution of new taxa, and (4) potential areas for development of biogeographically distinct island populations. R. B. Neuman (1984) recognized the midoceanic and marginal islands (some arc-related) as having played important parts in the development of the two biogeographically distinct Iapetus faunas, and these patterns have more recently been confirmed by Harper et al. (1996). First, the Early to early Mid Ordovician Celtic faunas developed in the intermediate to high latitudes of (1) island and microcontinental sites of Avalonia (to the south side of Iapetus) after it had split off from Gondwana in the Late Cambrian or Early Ordovician and (2) arc settings in marginal and intraoceanic sites (between the southern margins and the middle of the Iapetus Ocean), but these became accreted to Laurentia later in the Ordovician. Second, a contrasting, low-latitude Early Ordovician fauna (Toquima–Table Head assemblage) developed on the northern side of Iapetus, in Laurentian marginal sites (including west-central Ireland and Norway).

15

16

 .  Harper and Mac Niocaill (2002) described these various marginal and intraoceanic Iapetus sites as acting, alternately, as “cradles” that provided sources for radiations onto adjacent platforms and “museums” (or refugia) for the otherwise relict taxa. Island sites in the Late Ordovician of eastern Australia (Macquarie Arc) have also acted as cradles for the dramatic radiations of siliceous sponges in the upper slope (Rigby and Webby 1988; Percival and Webby 1996) and the heliolitine corals in level-bottom areas of the midshelf (Webby and Kruse 1984). There was also ecologic differentiation of trimerellid brachiopod shell beds in the inner shelf (Webby and Percival 1983) and tetradiid coral biostromes and shoals in the midshelf (Webby et al. 1997). The wholesale recycling of Paleozoic oceanic crust in subduction-related plate tectonics has left a patchy record of preserved ocean-sourced fossiliferous rocks, the sequences commonly disrupted, frequently limited in continuity across time and space, and sometimes metamorphosed. In contrast, the stable continental platforms typically have accessible, temporally and spatially continuous fossiliferous successions that may exhibit well-preserved communities across a wide range of habitats and a fossil record that has largely survived intact. Consequently, the fossil record as reflected by our global biodiversity databases does not provide a balanced view of the diversification of Ordovician marine life. We need to find some way to correct for the preservational bias that exists between the more complete biotal records in continental platform deposits and the incomplete and fragmentary records of biotas from oceanic sites. Other Preservation and Sampling Biases Considerable efforts have also been made in this biodiversity study to minimize potential biases, especially in the areas of evaluating taxa, establishing a refined timescale, and standardizing the diversity measures, as already stated. Other sampling biases have significantly important implications for Ordovician diversity studies across global and regional scales (Miller and Foote 1996; Holland 1997; Patzkowsky and Holland 1999; Alroy et al. 2001). Miller and Foote (1996) raised the possibility that the great proliferation of familial (and generic) diversity through Ordovician time, as illustrated, for example, by Sep-

koski (1981a, 1995; see figure 1.1) using raw data, might be affected by sampling bias. They depicted analyses showing raw and rarefied data (the latter to allow for the uneven sizes of the coarse, British-based series sampling intervals used by them) for their nearly global-scale sample of more than 6,570 Ordovician occurrences (trilobite, brachiopod, and benthic mollusk genera). The data are from Laurentia, Avalonia, Bohemia, Baltoscandia, North China, South China, and, to a limited extent, Australia. The diversity curve based on the raw data showed an upward rise during the Late Ordovician (Caradoc), in contrast to the “sample standardized” rarefied plot, which exhibits a flattened, plateaulike trajectory of stabilized diversity through the Late Ordovician. At the regional level, sequence stratigraphy architecture has been shown to exert a primary control on the fossil record (Holland 1995, 2000; Holland and Patzkowsky 1999). It not only has a significant role for use in establishing global correlations and age relationships but also has important implications for biodiversity studies of at least a regional, basinwide scale. Holland (2000) predicted from sequence stratigraphic modeling (with verification of some examples from the field) that four types of biases occur within the sequence packages: sampling biases (rare, short range forms especially likely to be missing), facies biases (concentrations of first and last appearances at abrupt facies changes such as flooding surfaces), unconformity biases (concentrations of first and last appearances at sequence boundaries), and condensation biases (clustering of occurrences during slow sedimentation). Some of these biases have been recognized by Patzkowsky and Holland (1996, 1997, 1999) and Holland (1997) in their full faunal analyses of the biotas of the Upper Ordovician (Mohawkian-Cincinnatian) depositional sequences in the eastern United States. These studies demonstrate how sequence stratigraphic architectures bias the fossil record, especially the representation of fossil range data. This in turn affects interpretations of patterns of biodiversity change in such deposits. All such successions therefore need full stratigraphic analyses to establish whether the biodiversity results reflect true biologic change or merely relate to apparent diversity change, driven or controlled by sequence stratigraphic architectures. Smith (2001) also emphasized the important role of sequence

Introduction stratigraphic architectures, and also major transgressiveregressive cycles, in controling Paleozoic faunal diversification. Holland and Patzkosky’s insights demonstrate that, in future biodiversity studies, paleontologists should, if possible, work in close cooperation with sequence stratigraphers. Alroy et al. (2001) adopted a new database program of compilation and analysis with a variety of sampling and analytical protocols to minimize sampling bias for translating the sampled fossil occurrence data into global Phanerozoic marine diversity curves. It aims to standardize sampling levels and to define counts of taxa. Eventually this major database initiative is intended to cover the entire Phanerozoic record of terrestrial and marine fossils across all geographic regions, but currently the sampling focuses only on well-studied regions through two halves of the Phanerozoic: interval 1 from the base of the Upper Ordovician (“Llandeilo”) through Carboniferous (apparently mainly across the Paleozoic diversity plateau), and interval 2 from middle Jurassic through Oligocene. The core taxonomic groups for interval 1 analysis include anthozoans, brachiopods, echinoderms, mollusks, and trilobites, with the current compilation dominated by inclusion of data from North American and European localities. Sampling of bryozoans, conodonts, and graptolites is as yet inadequate. Interval 1 has 15 roughly equal bins of 10.7 m.y., so the Late Ordovician Epoch spans only about two or three sampling bins. From an Ordovician perspective, not only is this a rather limited span of genus-level sampling (only depicting the last part of the Ordovician Radiation), but the primary data have significant biases—the core groups are not completely representative, given that all the groups are benthic to the exclusion of pelagics and they are geographically restricted in North American and European sites that, during the Late Ordovician, predominantly occupied low paleolatitude positions. The establishment of standardized analytical protocols to minimize sampling biases for calculating diversity estimates of previously collected data is indeed a laudable development but no real substitute for maintaining a wide array of intensive field- and laboratory-based biodiversity studies—especially regional and global investigations that fully document and sample biotas from all habitats (where applicable, making allowances for sequence stratigraphy architectures) and across all

paleolatitudes, tied to highly calibrated and resolved timescales. Timing of Major Radiations of Level-Bottom, Reef, and Infaunal Communities Level-Bottom Communities. The timing of the major Ordovician radiations of the level-bottom and the reef communities proves to be markedly different (Webby 2002). The best records of these faunal changes are preserved on the North American (Laurentian) platform, and they seem to be decoupled from one another (cf. Sheehan 1985). The significant radiation event that relates to the appearance of the level-bottom community is recorded from the base, into the lowest part, of the Middle Ordovician (lower Whiterockian, i.e., TS.3a–b; see table 1.3). The base of the Whiterockian represents the boundary between Boucot’s (1983) level-bottom community groups, Ecologic-Evolutionary Units (EEUs) III and IV— typically units lasting tens of millions of years with community stability (stasis) bounded by pulses of extinction (Brett 1995). These Ordovician units were renumbered EEUs P1 and P2 by Sheehan (1996, 2001a) on the basis that they were equivalent to the first two subdivisions of the Paleozoic EF. Droser and Sheehan (1997) noted that, because EEUs P1 and P2 are associated with the Ordovician radiations, they are not typical. The community complexity increased during EEU P1 because members of the Paleozoic EF were being added while representatives of the Cambrian EF were still in place. By the beginning of EEU P2, however, the community ecology of the Paleozoic EF was “fully in place” (Sheehan 2001a:244). The early Mid Ordovician radiation is characterized by both taxonomic and ecologic diversification changes that are especially well documented in the Great Basin successions of the western United States (Droser and Sheehan 1995, 1997b; Droser et al. 1996). The features include major turnover of faunas, changes in shell-bed composition, and hardground evolution. The shell-bed concentrations were dominated by brachiopods and ostracodes, instead of the previously important trilobite-dominated shell beds (Droser et al. 1996). The trilobites (representatives of the Whiterock Fauna; Adrain et al. 1998; see also chapter 24) diversified significantly, but the group no longer had a prominent role in shell-bed development

17

18

 .  —that is, the taxonomic and ecologic diversifications had effectively become decoupled. Bottjer et al. (2001) have regarded the early Mid Ordovician radiation event as exhibiting signals of second-, third-, and fourth-level ecologic changes, which reflect shifts ranging from major structural change in the ecosystem to minor community-level changes. Kanygin (2001) outlined the Ordovician diversification of biotas in the Siberian Platform and the Verkhoyansk-Chukchi fold-belt region of northeastern Siberia, highlighting a dramatic diversity change in the level-bottom communities of these successions across virtually all biotal components, apparently close to the base of the Llanvirn (i.e., mid Darriwilian— TS.4b). Both taxonomic and ecologic changes were involved. The Early to early Mid Ordovician (TS.1a– 4a) biotas were reported to be closer in aspect to Cambrian assemblages than to the markedly diverse mid Darriwilian to Late Ordovician record (Kanygin 2001). The Early Ordovician platform succession is dolomitic and dominated by occurrences of stromatolites. Trilobites are the most common faunal components, with brachiopods, cephalopods, conodonts, “monoplacophorans,” and sponges less common, and ecologic relationships remained generalized. In contrast, the record above the base of the Llanvirn has sudden appearances of abundant bryozoans and ostracodes, some crinoids, acritarchs, and chitinozoans, as well as continuation of moderately common trilobites, brachiopods, and conodonts. Most of the benthic species were endemics, and a host of ecologic innovations was also reported—for example, development of short-ranging benthic and “subpelagic” ostracodes (the latter adapted to passive hovering and active swimming) that densely occupied both shallowshelf and deep-shelf habitats. This may prove to be a fundamentally wide-ranging, coupled taxonomic and ecologic diversification event that occurred long after the ecologic diversifications in “level-bottom” communities of the early Whiterockian (TS.3a–b) in North America. However, the fact that this dramatic diversity change is associated with a regional deepening event, involving a sudden shift from restricted conditions (presence of dolomitic and stromatolitic deposits) to more normal open marine conditions, suggests that these Siberian successions require fuller stratigraphic analyses to determine whether the fossil records represent true or apparent biodiversity change.

Reef Communities. Modern coral reefs have the greatest levels of known diversity and productivity in marine ecosystems, and there seems little reason to doubt that Ordovician reefs, at least the metazoan/ calcified algal-dominated structures, relative to other Ordovician communities, were equally richly diverse and productive. Webby (2002) established that the distribution of Ordovician reefs, prior to the Mid to Late Ordovician boundary and the first appearances of well-skeletonized metazoan- and algal-dominated reef communities, included mainly microbial-dominated reef structures. Prokaryotic, photosynthetic microorganisms (mainly blue-green “algae” or cyanobacteria) were ultimately responsible for establishing the microbial communities that precipitated the various types of stromatolites, thrombolites, mats, and films and contributed the calcified microbes, such as Girvanella, Renalcis, and Epiphyton, to the initial Ordovician reef ecosystem. Sometimes these microbial constructions were associated in consortia with subordinate receptaculitids (Calathium), lithistid sponges (Archaeoscyphia), stromatoporoids (Pulchrilamina), bryozoans (Batostoma), and even rarer, tabulate corals (Lichenaria). Some structures, especially in the late Tremadocian, developed into huge, kilometer-scale barrier reef complexes along parts of the Laurentian shelf margin, and other large (near 100 m high) stromatactis-bearing carbonate mud mounds formed in parts of the outer platform and slope. These patterns were maintained through the Early Ordovician to the mid Darriwilian, until the dramatic change in the latest Darriwilian when the new, well-calcified metazoanand algal-dominated reef ecosystem emerged. The earliest stage of the great metazoan/algal reef expansion is best exhibited in on-shelf sites through the Chazyan to early Mohawkian interval (TS.4c–5b) of eastern North America (Webby 2002: figure 6). Stromatoporoids, tabulate corals, bryozoans, lithistid sponges, pelmatozoan echinoderms, and solenoporan algae established themselves during the late Darriwilian (Chazyan) in an array of new and complex community interrelationships, dependent on their differing frame building, encrusting, and sedimentproducing roles. Taxonomic diversification involved the bryozoans, lithistid sponges, a few stromatoporoids, and a few tabulate corals. Framework cavities became colonized by three types of cryptic niche dwellers: a first indubitable colonial organism (bry-

Introduction ozoan Batostoma), as well as taxa representing bioeroders and bioturbators. Microbial components such as localized stromatolites and Girvanella crusts continued to be associated, but they were now subordinate in the reefs. The second stage of reef development during the Caradoc (early Mohawkian) involved expansion of bryozoan-dominated reefs into the offshore—shelf-edge to downslope—habitats. The record is again most complete in eastern North American sites. Some of the more massive, composite structures were up to 50 km across and 80 m high near the shelf edge. By the mid Caradoc (TS.5c), the reefs had spread circumglobally to Greenland, Baltoscandia, Arctic Russia and the Urals, Kazakhstan, China, and Australia, to on-shelf sites near continental margins, and to island arcs. Another significant event is Wilson and Palmer’s (2001a) Ordovician “bioerosion revolution” (see discussion in chapter 34), involving a marked increase in the numbers of bioeroding organisms actively colonizing reefs and hardgrounds. This bioerosion event seems to be strictly correlative with the initial, late Darriwilian expansion of metazoan and calcified algal reefs. Infaunal Communities. A major increase in the intensity and depth of bioturbation has been recognized by Droser and Bottjer (1989) in the infaunal records of shallow-water carbonate successions of the Great Basin, western United States. This change marks a major infaunal biodiversity event, signifying an increased utilization of infaunal ecospace. It presumably reflects a restructuring of the infaunal ecosystem (at least in shallow-water carbonate platform habitats) and probably represents the evolutionary development of new and markedly different softbottom burrowers. It is highlighted by the morphological changes of one particularly common trace fossil (a complex branching burrow) called Thalassinoides, which in its Lower to Middle Ordovician record forms mazelike structures, which are replaced in the Upper Ordovician record by more complex boxwork structures to a depth of 0.3 to 1 m. The precise timing of this intensification of bioturbation, with the accompanying morphological change from the mazelike to boxwork structures, has not been determined. Unfortunately, much of the upper Darriwilian to middle Caradoc succession is not in the

typical “carbonate inner shelf ” facies that contains the key ichnofossil—either it is in a coarse nearshore quartzite and sandstone sequence, or the record is missing. Consequently, it is impossible at the present time to determine precisely when, within that undiagnostic interval representing about 15 m.y., the bioturbation change occurred and whether it was sudden or gradual. It may represent an infaunal radiation event of late Darriwilian (TS.4c) age—more or less contemporaneous with the initial expansion of metazoan and calcified algal reefs—or later, at some time during the Late Ordovician (Caradoc—TS.5a–d). Sections with a more complete Middle to Upper Ordovician stratigraphic record in the “carbonate inner shelf ” facies need to found elsewhere in North America and/or in other continental platform successions so that this infaunal event can be more precisely defined and correlated.

Ordovician Radiation in the Pelagic Realm Pelagic organisms—broadly divided into planktic (suspended, floating, and/or drifting) and nektic (independently mobile, swimming) types—lived in waters of the open ocean, as well as the waters overlying the continental shelves (and platforms). Generalized aspects of the origins and early diversification of pelagic faunas have been outlined by Rickards (1990), Signor and Vermeij (1994), and Rigby (1997). Rigby (1997), for example, has argued that all the main planktic groups had benthic origins—indeed, the plankton and benthos were inextricably linked (Signor and Vermeij 1994). Apart from the graptolites, few preserved, predominantly Ordovician pelagic groups have been studied intensively enough to understand how they became ecologically differentiated (Rickards 1975; Cooper et al. 1991; Underwood 1993; Finney and Berry 1997; Cooper 1999c). Furthermore, rather limited detailed attention has yet been given to studying the associated deep-water sediments, for example, as represented by biogenic siliceous sediments (radiolarian oozes) and organic-rich black shales (graptoliteand/or unicellular algal-dominated). Preservational biases were probably greater for pelagic forms than for benthos, given their exposure to predation, dissolution, or decay in their long descents through the water column before accumulation in sediments on the seafloor (Butterfield 1997). However, to some

19

20

 .  extent this would have been offset by a more continuous record of deep-water sedimentation and by the larger population sizes, especially in plankton blooms, which would have increased the chances of a few skeletons becoming entombed and preserved in bottom sediments. The preserved Ordovician record includes the planktic groups (unicellular green algae, acritarchs, chitinozoans, graptolites, radiolarians) and the mainly nektic groups (some conodonts, some nautiloid cephalopods, a few trilobites, and a few phyllocarid crustaceans). A few representatives of the conulariids and the polychaetes (scolecodonts) also may have had pelagic life habits, but there is insufficient evidence to review them meaningfully at this stage. Probably only a very small percentage (perhaps less than 10 percent) of the diversity of organisms that inhabited the open oceans in the Ordovician is likely to have been preserved—the bulk were soft-bodied organisms that left no trace. In addition, the overwhelming majority of dispersive planktic larval stages of Ordovician metazoan phyla, including most of the skeletonized benthic marine animals, are unlikely to have been preserved. These would have formed a very large component of the plankton in the water column—a very high diversity of soft-bodied larvae (Signor and Vermeij 1994)—that would have been most common closer to coastal areas. Acritarchs, Unicellular Green Algae, and Chitinozoans Acritarchs and unicellular green algae were the main preserved microplanktic groups of primary producers in the Ordovician oceans. The acritarchs exhibit a broad twofold “provincial” differentiation across paleolatitudes (chapter 32), and high levels of diversification were maintained (Tappan and Loeblich 1973) in one region or another through all intervals of Ordovician time. In addition, levels of productivity were high, with samples of more than 100,000 specimens per gram of rock recorded in certain offshore sites (Dorning 1999). The highest acritarch productivity is usually in the upper offshore, much reduced in the lower offshore, and mainly replaced by leiospheres (featureless sphaeromorph members of unicellular green algae) in more distal sediments.

The chitinozoans are presumed to be eggs of unknown soft-bodied metazoans and probably also a part of the plankton (Paris and Nõlvak 1999; chapter 28). A similar broadly twofold provincial differentiation of higher-paleolatitude and lower-paleolatitude assemblages was maintained (Achab 1991) and abundances of up to a few thousand specimens per gram of rock in the higher paleolatitude samples, compared with a few tens from equivalent low paleolatitude samples (chapter 28). From their initial appearances in the Early Ordovician (early Tremadocian), the chitinozoans expanded steadily to a first diversity peak in the Darriwilian and then, after a short-lived lowering, to a second, comparable diversity pulse during the early Late Ordovician (mid to late Caradoc). Graptolites Rickards (1975, 1990) considered the graptolites to be probably the first abundant macroplankton in marine habitats, with their food supply mainly minute phytoplankton and zooplankton. According to Cooper (1999c), the planktic graptolites probably arose from a sessile dendroid-type ancestor that occupied a deeper-water, outer-shelf site close to the shelf-slope margin in the earliest Ordovician (early Tremadocian). The major evolutionary event is thought by Cooper to have involved the larval stage acquiring the ability to initiate its skeletal growth without first attaching to the substrate. What followed was the major Tremadocian expansion of these earliest planktic graptolites into onshore (shelf ) and offshore (oceanic) habitats, with vertical differentiation into the shallow-water epipelagic depth “zone” and the deeper-water mesopelagic and bathypelagic depth “zones.” The boundaries between epipelagicmesopelagic and mesopelagic-bathypelagic “zones” were placed at depths of 200 m and 1,000 m, respectively (Cooper et al. 1991). The vertical depth-related differentiation of graptolite assemblages was followed during the late Early Ordovician to Darriwilian (Arenig-Llanvirn) interval by the equally important lateral, “provincial” differentiation across latitudes and the greatest radiation of graptolite species—with a spectacular diversity spike in the early Arenig representing dichograptid appearances (Cooper 1999c: figure 1).

Introduction Radiolarians The radiolarians utilized seawater saturated with silica for their skeleton formation and were consequently major contributors to the formation of Ordovician cherts. They therefore had an important role in establishing sinks of oceanic silica in some deeper basinal habitats (Malvina et al. 1990). Ordovician radiolarian chert localities are widely distributed in Kazakhstan, New South Wales (Australia), Newfoundland, Nevada, Scotland, Spitsbergen, and Gansu (North Central China). Although Nazarov and Ormiston (1985) originally claimed that spherical-type zooplanktic radiolarians appeared in abundance for the first time during the Early Ordovician, recent reports of assemblages in the mid to late Late Cambrian of Newfoundland (Won and Iams 2002), the Siberian Altai Mountains, and central Kazakhstan (Tolmacheva et al. 2001a) now suggest that their initial major diversification into continental slope and deep basinal habitats may have commenced slightly earlier. Similar assemblages are reported from the early Tremadocian (chapter 11) but are then replaced in successive pulses of early Arenig and Mid Ordovician diversity increase by more typical, Ordovician-type associations. Conodonts Sweet (1988b) interpreted conodonts as pelagic, probably nektic, marine animals—a swimming predator that occupied a broad range of geographic and environmental habitats. Two main ecologic models were originally proposed to explain the distribution patterns: (1) Seddon and Sweet (1971) favored a depth-stratified model that was more or less based on the assumption that conodonts were pelagic organisms; and (2) Barnes and Fåhraeus (1975) preferred a biofacies-controlled model, related to the lateral segregation of predominantly nektobenthic organisms and some benthics—only simple coniform taxa being interpreted as pelagics. Many refinements and modifications to the two basic “pelagic” and “nektobenthic” conodont schemes have since been proposed (see review in Pohler and Barnes 1990), and more generally applicable fossil distributional modeling applications have also been outlined by Tipper (1980). Some of the taxa recorded as from “deep/cold environments” in Albanesi and Bergström’s analysis (chapter 29) are pelagic forms.

Many detailed studies of conodont biofacies have been undertaken over the past two decades, especially in North America (e.g., Sweet and Bergström 1984; Ji and Barnes 1994a; Pohler 1994; Johnston and Barnes 1999), but Rasmussen’s (1998) detailed, computer-based cluster analysis of mid Darriwilian conodont faunas was the first to demonstrate that the deep-water Protopanderodus-Periodon biofacies of the North Atlantic provincial realm differed from other biofacies in exhibiting exactly the same associations of cosmopolitan forms (including coniform taxa) in both Baltica and Laurentia (specifically eastern Laurentia). This was despite the fact that these continental blocks occupied different sides of the Iapetus Ocean and lay in different paleolatitudes during Mid Ordovician time. When these cosmopolitan taxa were removed from the biogeographic analysis of the eastern Laurentia fauna, the region became much more provincially similar to the North American Midcontinent provincial realm than was previously shown. His study established the deep-water ProtopanderodusPeriodon association as an oceanic biofacies, implying that the conodonts were pelagic forms. Armstrong and Owen (2002a, 2002b) provided further insights on how the conodont biofacies (and biodiversity) became differentiated, with the shelf biofacies composed of nektobenthic forms, and the oceanic biofacies of wide-ranging pelagic taxa, recognizing that they lived in shallower, stratified oceanic water masses. Their presence-absence data for conodont genera were compiled from sections across onshore-offshore profiles in Laurentia and Avalonia (to either side of the Iapetus Ocean) for three Mid to Late Ordovician “time-slice” intervals—late Darriwilian– early Caradoc (TS.4c–5a), late Caradoc (TS.5d), and mid to late Ashgill (Rawtheyan, TS.6b)—with all genera coded to a particular biofacies. Three depthrelated oceanic biofacies and three (usually two) laterally segrated shelf biofacies were recognized. The shelf biofacies were regionally distributed, ranging from onshore to offshore and terminating offshore. The oceanic biofacies included faunal elements usually regarded as “North Atlantic” provincial components and many coniform genera. It had a spread from the inner to outer shelf that was independent of benthic biofacies. Higher diversities were associated with the inner shelf and the upper water column in both Laurentia and Avalonia.

21

22

 .  Armstrong and Owen (2002a) stressed the importance of defining the biofacies in order to reconstruct the biofacies architecture of the regions and as a basis for establishing meaningful analysis of paleogeographic relationships. Clearly many previous studies of provinciality are now outdated, as they employ admixtures of shelf and oceanic taxa. The oceanic biofacies reflected water-mass structure, and this was likely to have caused different ecologic constraints and dispersal mechanisms for oceanic taxa, as compared with time-correlative benthic taxa. Other complications may apply where changes in temperature and/or density cause vertical movements of the thermally stratified water masses, such as in upwelling zones at the continental margins. Armstrong and Owen’s biofacies insights have important implications for biodiversity studies. In the future the biofacies records must be dissected into their benthic and pelagic components so that more meaningful patterns of biodiversity change can be determined for each very different ecologic realm. Zhen and Percival (2002), working in the Australasian region, have independently recognized the difficulties of utilizing the biogeographic subdivisions based on the Midcontinent and North Atlantic provincial realms (or provinces). They have adopted a global approach that to some extent overlaps with the new initiatives of Rasmussen, Armstrong, and Owen, arguing that most of the organisms occupying the shelf regions were nektobenthic or benthic (representatives of their “Shallow-Sea Realm” [SSR]) and recognizing a separate, important component of pelagic forms that have cosmopolitan faunal relationships and include coniform taxa (their “Open-Sea Realm” [OSR]). Temperature, depth, and salinity are regarded as controlling influences, though endemic forms show more restricted habitat preferences in the SSR. The SSR is subdivided into six provinces: Australia, Laurentia, North China (Tropical Domain), South China and Argentine Precordillera (Temperate Domain), and Baltoscandia (Temperate and Cold domains). The OSR (apart from a tropical domain) remains undivided. Zhen and Percival (2002) have recommended that usage of the traditional North American Midcontinent conodont “province” now be strictly equated with the low-latitude domain (Laurentia) of their SSR. They further argue that the North Atlantic conodont “province,” which rep-

resents a composite of temperate and cold domains (Baltoscandia) in the SSR and an as yet undefined domain of the OSR, should now be abandoned. In terms of Ordovician biodiversity relationships, the assemblages in Zhen and Percival’s OSR are significantly pelagic biotas. Armstrong (1997) recognized a deep or cool, predominantly shelf and slope assemblage of probable nektic taxa from the Southern Uplands of Scotland that he referred to the “Periodon-Pygodus Restricted Species Association” (sensu Bergström and Carnes 1976). This association of pandemic species, from the Pygodus anserinus Zone (latest Darriwilian–earliest Caradoc), is identical to correlatives in eastern North America (Newfoundland and Tennessee) and Baltoscandia (i.e., clearly a part of the oceanic biofacies). Armstrong (1996) had previously interpreted the patterns of distribution of conodonts across the Ordovician-Silurian boundary in terms of the changing ocean state, presenting a broad outline of biofacies profiles from the late Caradoc onward. This included reference to Sweet and Bergström’s (1984) biofacies data and to the deepest association from basinal cherts, called the Dapsilodus-Periodon biofacies (see further discussion of the chert association later in this chapter). Sweet and Bergström (1984) recognized this biofacies as the deepest they encountered in their survey of the late Caradoc–earliest Ashgill interval. It is preserved with graptolites and radiolarians in the Ouachita trough—a rifted reentrant into the Laurentian continental interior of Oklahoma that lay in low paleolatitudes. Tolmacheva et al. (2001a) described a remarkable 35-m-thick biogenic accumulation of ribbon-banded radiolarian cherts and shales with a nearly continuous zonal record of conodont assemblages in a composite section of the Burubaital Formation, central Kazakhstan. These range from Late Cambrian (with continuity across the Cambrian-Ordovician boundary), through the Tremadocian, to the early Arenig. The assemblages were considered to represent pelagic components of the oceanic biofacies, possibly associated with an area of equatorial upwelling. Earlier, Dubinina (1991) and Popov and Tolmacheva (1995) outlined the conodont succession in sections in central Kazakhstan, at Sarykum and Burubaital (about 240 km apart), respectively, through essentially the same interval of chert-dominated sediments and re-

Introduction vealing essentially similar temporal patterns of faunal change. In the lower part of the succession, that is, through the Late Cambrian (Eoconodontus notchpeakensis Zone) to Tremadocian (lower Drepanoisodus deltifer Zone) interval, the associations were dominated by two intermixed faunal components, the mainly undescribed, very simple, conelike protoconodonts and paraconodonts and the true conodonts (euconodonts)—including many diagnostic zonal species of Cordylodus, Eoconodontus, Hirsutodontus, Cambrooistodus, and Teridontus. The intermixture of components was explained by Dubinina (1991) and Popov and Tolmacheva (1995) as due to thermally stratified water masses—the protoconodont and paraconodont elements representing the cooler association that occupied waters below the thermocline, while the component of euconodonts lived in warmer, shallower waters above the thermocline. During the D. deltifer Zone (early to mid Tremadocian), however, a significant faunal change occurred in the Kazakhstan sections, interpreted by Popov and Tolmacheva as indicating the disappearance of the well-defined, two-layered stratification. The warmer association disappeared, and the protocondont and paraconodont assemblages were replaced by a single, low-diversity euconodont assemblage. This association continued to be represented in Kazakhstan from the D. deltifer Zone, through Paraoistodus proteus and Prioniodus elegans, to Oepikodus evae zones (mid Tremadocian–mid Arenig). This major oceanic event in the D. deltifer Zone also seems to equate with emergence of well-marked Ordovician conodont provincialism in the Ordovician, especially in the SSR. Two less-pronounced Late Cambrian phases of provincialism have been identified by Miller (1984), the first with the appearance of differentiated pelagic euconodonts in warmer waters of the late Franconian (mid Late Cambrian), and the second when the species of the mainly pelagic Cordylodus and the questionably nektobenthic genera of the Teridontus lineage invaded shallow, low-latitude areas in the latest Cambrian (base of Ibexian). Trilobites Most Ordovician trilobites lived as benthic or nektobenthic organisms in reasonably close contact with the seafloor, though, of course, like most meta-

zoan phyla, they probably had dispersive planktic larval stages. A few small trilobite groups, however, developed active swimming roles in a pelagic adult life (Fortey 1985; McCormick and Fortey 1998, 1999; see also Fortey in chapter 24). Fortey (1985), using three main criteria—their functional design, nature of geologic association, and analogy with living, large-eyed, ocean-dwelling isopod and amphipod crustaceans—determined that two main groups of pelagic trilobites were initially differentiated early in Ordovician time. The first was mainly based on the family Telephinidae, with characteristic Carolinites, Opipeuter, and four other genera—forms that were regarded by Fortey (1985) as having large bulbous eyes and comparatively poorly streamlined, smaller body forms. They have been interpreted as becoming well adapted to active surface swimming in near equatorial waters, constrained between 30 degrees north and south paleolatitude and in the upper part (epipelagic “zone”) of the water column, being a widespread community across all biofacies (McCormick and Fortey 1998, 1999). Elements of this epipelagic community maintained a pan-tropical span from early Arenig through Mid Ordovician time, but Telephina and Phorocephala dispersed to higher paleolatitudes during the Late Ordovician, prior to their end Ordovician demise. The second group was characterized by representatives of the family Cyclopygidae—mainly large-eyed genera such as Pricyclopyge, Cyclopyge, Degamella, and perhaps up to seven others—as well as other, rather aberrant forms having possible remopleuridioid affinities (Bohemilla and Cremastoglottos). Some of these, such as Cyclopyge and Bohemilla, were smaller, poorly streamlined, sluggish swimmers and others, such as Degamella, larger and streamlined, more efficient swimmers. Given these differences, Fortey and Owens (1999) suggested that the two groups may have fed on different varieties of plankton or that the larger forms may have preyed on smaller plankton feeders such as the phyllocarid crustaceans (see later in this chapter). Their overall distribution was markedly different from that of the telephinids, characterized by the cyclopygid biofacies that occupied the mesopelagic “zone” (probably below 200 m water depth), offshore from marginal areas across moderately high paleolatitude regions of peri-Gondwana, mainly between western and central Europe (Britain,

23

24

 .  France), to Kazakhstan and China (McCormick and Fortey 1998). This pelagic cyclopygid biofacies was long ranging and conservative, first developing in the Tremadocian of Argentina, becoming more widespread through late Early Ordovician (Arenig) to Caradoc time, then being more restricted within low paleolatitudes during the Ashgill (Scotland, Quebec), before its extinction at the end of the Ordovician (Fortey in chapter 24). Phyllocarid Crustaceans Although many phyllocarid crustaceans appear to have maintained benthic or nektobenthic habits (Rolfe 1969), a few distinctive Ordovician representatives of the group, in particular the widely distributed genus Caryocaris (Racheboeuff in chapter 25), adopted a truly pelagic mode of life, with preservation in deep-water graptolitic shale deposits of offshore, mainly slope to basinal, sites. The morphology of Caryocaris was characterized by thin carapaces and expanded, leaflike furcal rami that would have allowed it to adopt a free-swimming habit in the water column. In places, large concentrations of specimens may be preserved on bedding planes that perhaps suggest that they also lived in schools at a particular depth within the water column. They have a wide distribution across high to low paleolatitudes and a long Ordovician record of occurrences, from the early Tremadocian to Caradoc. Fortey (1985) has noted that phyllocarids were common in the deeper, pelagic cyclopygid trilobite biofacies and that they may have been preyed on by larger, streamlined trilobite species of that biofacies. In comparing the Ordovician species of Caryocaris and related forms with certain freeswimming modern crustaceans, Vannier et al. (2003) recognized that in all probability the Ordovician “caryocaridids” lived in epipelagic to mesopelagic depth zones of the water column, especially “marginal” outer-shelf to slope settings. Nautiloid Cephalopods Late Cambrian nautiloid cephalopods (members of the order Ellesmereocerida) exhibit the initial development of a siphuncle (Yochelson et al. 1973), which allowed intercommunication through a strand of soft tissue for supply of blood and gaseous exchange to the closed off, septate, apical part of the

shell. This unique evolutionary innovation paved the way for all cephalopods to achieve very effective buoyancy control during their Ordovician and later development. The paleoecology of the Late Cambrian precursors is not well understood. According to Flower (1957), the known shells are small, have closely spaced septa, and are associated in limestones with benthic trilobites. Hence, they probably lived near or on the bottom. While the close spacing of septa in these early nautiloids may have been a simple device to reduce buoyancy in the shells (Teichert 1967), it is not known whether the intervening spaces (camerae) could yet be used for gaseous exchange. Holland (1987: figure 4) depicted how the two basic Cambrian types of short, nearly vertically oriented, slightly curved shells gave rise to the two great lines of nautiloid descent—coiled forms deriving from shells with an exogastric curve (ventral to the outer convex side) and the longicones evolving by straightening of shells with an endogastric curve (ventral to the inner concave side). The apical part of the longicone of later forms became partially filled by cameral and siphuncular deposits, enabling a horizontal life mode to be maintained. The greatest ever diversification of nautiloids occurred during the Ordovician, with all the major higher taxon groups appearing by the early Mid Ordovician (Flower 1976; Crick 1981; House 1988). The various representatives have been interpreted broadly as exhibiting a full range of benthic, nektobenthic, and pelagic life habits, but at the present time only a few can be assigned indubitably to the pelagic realm. No detailed systematic survey of Ordovician nautiloids has yet been attempted, comparable to that undertaken by Fortey (1985) to elucidate the life habits of trilobites. It seems that pelagic forms have been derived from the two main body plans mentioned earlier. In terms of the major Ordovician groups, the order Tarphycerida was dominantly coiled, with the discoidal forms becoming nimble and very maneuverable swimmers, while the orders Michelinocerida, Actinocerida, and Endocerida had mainly horizontally disposed longicones, though only the first two groups may have had roles as active swimmers. Probably the majority of the larger Ordovician longicones were carnivores (predators and scavengers) that occupied positions toward the top of the Ordovician food chain. With the ability to maintain precise

Introduction buoyancy controls and an efficient means of jet propulsion, both forward and backward, dependent on the way they oriented their hyponome, they would have been capable of maintaining an active swimming mode. The largest longiconic forms may, however, have had problems because of their size and weighting (Holland 1987), restricting them necessarily to a mainly nektobenthic existence near the seafloor. Examples include the giants referred to by Holland (1987) with estimated lengths of nearly 9 m (including a body chamber), based on estimates from a number of mainly incomplete endocerid longicones. The specimen from Watertown, New York (Flower 1957), was probably of early Late Ordovician age. Crick (1990:147–148) considered that Paleozoic nautiloids were representatives of the “shallow-shelf vagrant benthos,” not strictly a part of the nekton capable of oceanic dispersal. Consequently, they were capable of dispersing only “over shallow stretches of open ocean.” Stait et al. (1985), Webby (1985, 1987), Stait and Burrett (1987), and Percival in Webby et al. (2000) documented the biogeographic distribution of Ordovician nautiloids in the Australasian region, recognizing that nautiloids preferred the extensive shallow waters of carbonate platform areas. There they adopted nektobenthic habits, for example, as represented by the diverse and predominantly endemic Tasmanian faunas described by Stait (1988). However, in central New South Wales—originally a part of an island arc–related terrane, at least 1,000 km away from the cratonic East Gondwanan margin, including Tasmania (Webby and Percival in Webby et al. 2000)—the few nautiloids in the island shelf carbonates were longiconic orthoceratids and discoidally coiled tarphycerids, almost exclusively free-swimming, long-ranging, cosmopolitan forms. An additional occurrence of the wide-ranging, cosmopolitan genus Bactroceras, representing a specialized, surviving group of ellesmerocerids (chapter 21), is associated in deepwater graptolitic shales. It also appears to be a nektic component. Hewitt and Stait (1985) argued that the species, B. latisiphonatum, with its tubelike thickening of connecting rings that strengthened the septa, was able to withstand higher hydrostatic water pressures before imploding, compared with the extant Nautilus. Data from Westermann and Ward (1980) indicate that the depth of implosion of Nautilus shells is 600 m. Applying strength indices to the Bactroceras

data, Hewitt and Stait (1985) estimated a slightly greater implosion depth of about 800 m for the New South Wales species. Consequently, there seem to be two likely scenarios, given that the offshore New South Wales island arc was separated from other parts of the Gondwanan margin by the Wagga Sea (a marginal sea). If the nautiloids tried to cross the Wagga Sea as benthic or nektobenthic elements, maintaining close contact with the seafloor, they would have exceeded implosion depths of 800 m (using comparable depths across the present-day counterpart, the Tasman Sea). The alternative, and preferred, view is that the New South Wales nautiloid assemblages crossed the oceanic barrier by swimming (perhaps also in part drifting in ocean currents) but remaining in the epipelagic to upper mesopelagic zones of the water column during this oceanic dispersal. Hence the Late Ordovician Wagga Sea was a barrier to the Tasmanian benthic and nektobenthic nautiloids but not to wider-ranging nektic forms. Crick (1980) presented a global assessment of late Early Ordovician to early Darriwilian (Arenig) nautiloid biogeography and differentiated a number of groups that had higher mobility and a more cosmopolitan distribution, recognized as common in the oceanic environment (his “geosynclinal facies”). Longicones of the ellesmerocerid and protocycloceratid families (order Ellesmerocerida), the actinocerid family (order Actinocerida), and proterocamerocertatid and endocerid families (order Endocerida) are well represented, as well as coiled-shelled tarphyceratid and estonioceratid families (order Tarphycerida). It may be inferred, then, that by the early Arenig at least some representatives of these forms had adopted pelagic life habits, possibly mainly within the epipelagic zone. On the other hand, Teichert (1967) considered that all the Early Ordovician cephalopods with longicones probably remained as bottom dwellers, given that calcareous endosiphuncular deposits—weighting structures so important for buoyancy control—had not developed significantly prior to the beginning of the Mid Ordovician, though he regarded the coiled shells of tarphycerids as already having partially colonized the pelagic realm. In the major diversification of the Mid Ordovician nautiloids, three groups—the actinocerids, orthocerids, and ascocerids—seem to have adapted at least in part to living in the pelagic realm. Possibly, components

25

26

 .  of twofold differentiation of distinctive faunal provinces across higher and lower paleolatitudes during the Mid Ordovician, outlined by Frey et al. (chapter 21), also extended into the pelagic realm. The longiconic actinocerids and orthocerids are mediumto large-sized groups that developed a variety of weighting structures (especially endosiphuncular and cameral deposits) toward the apical end, helping to stabilize their orientation in a horizontal position, in their role as active swimmers. The orthocerids have shells characterized by well-developed cameral deposits, subcentrally placed, tubular siphuncles that may either be empty or sometimes associated with endosiphuncular deposits, as well as orthochoanitic septal necks, thin connecting rings, and pointed apices. The secretion of calcareous cameral deposits added weight to the apical (posterior) end as the shell grew, helping to maintain the horizontal life position in an equilibrium, with centers of gravity and buoyancy near one another and also acting to stabilize the shell with respect to its ventral and dorsal sides. Meanwhile, buoyancy control was maintained by gaseous exchange into the anterior camerae, effectively lightening the shell for active swimming (Flower 1976). The actinocerids, in contrast, feature a siphuncle with mainly cyrtochoanitic-type septal necks, an elaborate development of annulosiphonate, endosiphuncular deposits, and an associated vascular system that includes radial canals extending from the central tube. Larger siphuncles are usually placed toward the ventral side, though smaller siphuncles tend to be more subventral to subcentral. Cameral deposits also occur in many genera, and the apices are short and blunt. Both these longicone groups have long histories through the Paleozoic. The ascocerids were a small, comparatively poorly preserved, slightly cyrtoconic group, with thin, rather fragile shells, specialized septa that almost directly overlay the main body chamber, without cameral or siphuncular deposits, and an overall streamlined body design—well adapted to swimming in a horizontal orientation, presumably within the water column (Flower 1957; Furnish and Glenister 1964; see also Holland 1984: figure 1). They did not appear until the late Darriwilian (Chazyan) and then survived to the Late Silurian (Flower 1976). Nautiloid assemblages, as previously mentioned, sometimes have close associations with deeper-water

graptolite shales. Marek (1999) recorded a rich nautiloid assemblage from the Sˇárka Formation of the Prague Basin, Czech Republic (Bohemia), from a deeper-water black shale succession with rich accompanying biotas. Havlícˇek (1998:53), in outlining the stratigraphy of the Sˇárka Formation, gave special attention to the marked transgression and facies changes that accompanied the start of black shale deposition and the “sudden influx of new benthic and planktic biotas,” comprising graptolites, acritarchs, chitinozoans, conulariids, cystoids, crinoids, carpoids and other echinoderms, brachiopods, trilobites, gastropods, univalves, phyllocarids, and ostracods. The succession ranges through two graptolite zones: the Corymbograptus retroflexus and succeeding Didymograptus clavulus zones of mid to late Darriwilian (Llanvirn) age. Mergl (1999) outlined the Early to Mid Ordovician brachiopod community distribution for the Prague Basin, depicting the onshore-offshore community profile in time and place. As depicted (see Mergl’s figure reproduced in this volume— figure 17.10), the rich biotas mentioned here come mainly from the deeper-water environments of the Sˇárka Formation. Bohemia (part of the Perunican terrane; chapter 5: figure 5.1) lay off Gondwana in high paleolatitudes during Mid Ordovician time, forming part of the cooler “Mediterranean Province.” The trilobites comprise representatives of the mesopelagic community (cyclopygid biofacies), as well as some elements of the “atheloptic” assemblage—deep benthic forms with blind or much reduced eyes (Fortey in chapter 24). The nautiloid assemblage includes 20 species, all longicones, mainly orthocerids, and a few ellesmerocerids, endocerids, and an actinocerid as well (Marek 1999). It seems likely that this community was largely composed of nektic forms. Marek (1999) referred to the unusual richness of the nautiloid fauna as probably caused by a warming event (and related sea level rise mentioned by Havlícˇek 1998) and suggested that the nautiloids were deposited in depths no greater than 150 m. But the presence of the mesopelagic trilobite community and accompanying atheloptic elements suggests that depths may have been a little greater. Bogolepova (1999) briefly reviewed the Mid Ordovician nautiloid geographic distributions of the Mediterranean Province (Iberian, Armorican, and Bohemian regions), suggesting a strong invasion of elements, especially of ortho-

Introduction cerids and endocerids, from the Baltic (which lay in intermediate, ?warmer, paleolatitudes) to the “Mediterranean Province.” That mid Darriwilian invasion of more mobile swimming orthocerids and endocerids seems to be related to sea level rise and climatic amelioration, which accords with the patterns in the Prague Basin. It may represent the initial major colonization of higher paleolatitude parts of the pelagic realm. A few orthocerids from graptolitic shales (Utica Shale and equivalents) of Caradoc (late Mohawkian to early Cincinnatian) age across eastern and midcontinental North America have been reported as being relatively small and apparently thinner shelled, though larger specimens also occur (Flower 1957: 835). These may be pelagic nautiloids, but investigation of the mode of life of nautiloids preserved in black shales remains a challenge for future workers. ■

Ordovician World in Brief

The six short introductory essays outline the disposition of major terranes (chapter 5), the isotope patterns (chapter 6), the oceans and climate (chapter 7), the evidence for a possible Mid Ordovician superplume event (chapter 8), the end Ordovician glaciation (chapter 9), and the sea level changes (chapter 10). Cocks and Torsvik (chapter 5) provide one global and three South Polar Ordovician map reconstructions based on paleomagnetic, faunal, and facies data that show the positions of terranes (Gondwana, Laurentia, Baltica, and Siberia and many other blocks), the Panthalassic, Iapetus, Rheic, and Tornquist oceans, the equator, midlatitudes, and South Pole. The successive Ordovician map reconstructions (figures 5.1– 5.4) illustrate the changing patterns of displacement of continental blocks such as Baltica and Avalonia and the progressive closure of the Iapetus Ocean through Ordovician time. Island arcs are also identified. Shields and Veizer (chapter 6) look at the fluctuations of strontium, neodymium, carbon, oxygen, and sulfur isotope compositions in Ordovician seawater and how these records help in interpreting global dynamics and environmental change. Of particular significance is the sharp excursion in the strontium values across the Middle to Upper Ordovician boundary, mentioned by Barnes (chapter 8) as an indicator

of enhanced interaction by a possible mantle superplume with the hydrosphere. Barnes presents broad overviews of oceanographic and climatic events in the Ordovician world (chapter 7) and evidence for a mantle superplume event during the Mid Ordovician (chapter 8). The highlights are a world in a greenhouse state, most land located in the Southern Hemisphere, a complex spread of island arcs and microcontinents in the Iapetus Ocean, major phases of orogenesis and associated volcanism, a probable mantle superplume event, high sea levels, widespread epeiric seas, and end Ordovician glaciation. Barnes also identifies a number of events that may have been particularly important in shaping the overall patterns of Ordovician biodiversification. Brenchley (chapter 9) provides an outline of the main features of the short-lived end Ordovician (Hirnantian) glaciation in near polar areas of Gondwana and the timing of the associated climatic, isotopic, and biologic events. The rapid rate of environmental change is seen as being especially important in triggering the two phases of mass extinction, the first related to global cooling and the second associated with global warming with accompanying spread of anoxia. Nielsen (chapter 10) offers a new Ordovician sea level curve based on the Norwegian, Swedish, Estonian, and Latvian successions of Baltica, establishing close ties with parts of the Ross and Ross (1992, 1995) sea level curve of the North American (Laurentian) platform. Correlation of some 30 different sea level changes across the two regions suggests that they represent substantially eustatic fluctuations of a consistent, wideranging (possibly global) sequence stratigraphic framework. Relationships between the sea level changes in Baltoscandia and the global time slices (chapter 2) are shown in figure 10.3. Potentially, if the pattern of 30 short-term sea level changes can be matched on a truly global scale, then it will, especially if used in close conjunction with key biostratigraphic indicators and radiometric dates, provide a very good complementary basis for studying aspects of biodiversity change, especially species-level analyses that require the most finely resolved subdivision of Ordovician time. ■

Synopses of Fossil Groups

The individual topics in parts III (Taxonomic Groups) and IV (Aspects of the Ordovician Radiation)

27

28

 .  are presented in the following order: part III—the siliceous unicellular radiolarians (chapter 11), the main animal groups—sponges (chapter 12) to vertebrates (chapter 30), the problematic receptaculitids (of possible metazoan affinity) and the algae (chapter 31), the organic-walled, unicellular acritarchs (chapter 32), the miospore record and evidence for land plants (chapter 33); and part IV—the ichnofossil (trace fossil) record (chapter 34) and a new global synthesis (chapter 35). Noble and Danelian present the diversification history of radiolarians in chapter 11. This major siliceous, zooplanktic component of the Ordovician oceans shows a more or less progressive rise in genuslevel diversity through Ordovician time, to the mid Ashgill (table 1.3). The first assemblages in the Late Cambrian and Tremadocian differ markedly from typical Ordovician radiolarians. Ordovician-type groups with new body plans appear initially in the late Early Ordovician (early Arenig), and then successive familyand generic-level diversification pulses follow, during the Darriwilian and the Late Ordovician (mainly late Caradoc—TS.5d), leading to a diversity peak of 17 genera in the mid Ashgill (TS.6b). On the other hand, the species exhibit diversity maxima in the Darriwilian (TS.4b) and the early Caradoc (TS.5a). Carrera and Rigby (chapter 12), in their outline of species-level sponge diversification, depict a diversity curve with overall rise through Ordovician time but punctuated by increases to lesser peaks in the late Tremadocian and latest Early Ordovician (mid Arenig) and a more sharply peaked maximum in the mid Darriwilian due mainly to the radiation of orthocladine demosponges. Several higher-level sponge groups appeared next during the early Caradoc, contributing significantly to the overall Late Ordovician diversity rise, with culmination in a major early Ashgill (TS.6a) spike. This peak was followed by a dramatic mid Ashgill (TS.6b) decline. This decline occurs about one time slice earlier than the loss of diversity directly associated with the end Ordovician (TS.6c) mass extinctions, as seen in most other groups of organisms. Biogeographic and paleoecologic factors are shown to be influential in shaping the sponge diversification. In chapter 13, Webby records the genus- and species-level diversity patterns of the stromatoporoids, a group of calcified sponges that commonly contrib-

ute to the growth of reefs. Apart from a short-lived development of the problematic Early to early Mid Ordovician pulchrilaminids, the group has a mainly Darriwilian to Late Ordovician record. The specieslevel diversity curve appears steplike, with each diversity peak (successively the late Darriwilian, mid Caradoc, and mid Ashgill peaks) being higher than the preceding peak. Then a marked late Ashgill decline followed in the interval associated with the end Ordovician glaciation. Rates of genus-level origination were highest in the late Darriwilian and mid Caradoc, whereas the rates of origination at the species level were highest in the early Ashgill. Extinction rates at both the genus and species levels were the highest in the mid to late Ashgill. Van Iten and Vyhlasová (chapter 14) report that the conulariids had the highest ever generic-level diversities during the Darriwilian to mid Caradoc. That interval also exhibits the highest Ordovician specieslevel diversities. The conulariid records are from the low-latitude North American platform and the higherlatitude Prague Basin (Czech Republic). A relatively high proportion of genera are confined to either one region or the other. Only about one-third of the genera are common to the two regions. The authors also recognize a latitudinal diversity gradient with generic diversity levels declining toward the higher latitudes. Webby, Elias, Young, Neuman, and Kaljo (chapter 15) document a number of aspects of coral diversity in a general overview of coral groups, a global analysis of the tetradiid corals, and three regional analyses— on North American Late Ordovician (Cincinnatian) corals, Baltoscandian rugose corals, and Australasian corals, respectively. Of the two major groups, the tabulate corals appeared first in the earliest Ordovician (Tremadocian) but remained at a background level (one genus only) for nearly 25 m.y. (more than half the length of the Ordovician). Only then did they start to diversify through the late Darriwilian to Late Ordovician, becoming the dominant Ordovician coral group. In the Cincinnatian of Laurentia, for example, tabulates approximately outnumber the rugose corals by 2:1. In terms of the main higher-level diversification of the two groups, the tabulates diversified mainly between the late Darriwilian and the mid Caradoc, whereas the rugose corals first appeared during the early Caradoc, and by mid Caradoc time their main higher-level diversification had already occurred.

Introduction The tetradiid corals are a small, exclusively Ordovician group that are here separated from the tabulates. They first appeared and remained at background levels in the Darriwilian but then diversified rapidly during the early Caradoc, reaching a specieslevel peak of 17 species by the mid Caradoc. A small diversity decline followed, and then a second diversity peak was attained in the early to mid Ashgill, prior to their dramatic late Ashgill decline to end Ordovician extinction. The Baltoscandian rugose corals show a comparatively similar distribution pattern with more rapid diversification during the mid Caradoc and mid to late Ashgill against a background of generally rising diversity. Taylor and Ernst (chapter 16) provide a comprehensive analysis of the major radiations of bryozoans through Ordovician time, during which all Phanerozoic marine orders make their first appearances except one. The database comprises a total of 169 genera and 1,120 species. Patterns of diversity change and turnover show an exponential rise for both genusand species-level diversity from the Early Ordovician (mid Arenig) to Late Ordovician (mid Caradoc), with the most dramatic increase commencing in the Darriwilian. Then two sharp, steplike declines of species occur to the end Ordovician, whereas the genus-level diversity drop commences later with a more dramatic, mid to late Ashgill fall. This loss of genus-level diversity is considered to be related to Lazarus taxa and a poor fossil record rather than to extinction. All eight major groups diversified most markedly through the Darriwilian to Caradoc interval, though two showed significant increase earlier, and others attained their diversity peaks only later, in the early to mid Ashgill. Provinciality was probably greatest in the Arenig, declining progressively to the end of the Caradoc and then increasing in the latest Ordovician. A genus-level diversity survey of the brachiopods is presented by Harper, Cocks, Popov, Sheehan, Bassett, Copper, Holmer, Jin, and Rong (chapter 17). Formerly recognized as comprising only two broad divisions (“inarticulates” and “articulates”), the phylum is now grouped in three subphyla: the linguliformeans, craniiformeans, and rhynchonelliformeans. Each group shows markedly different taxonomic diversification histories. Most of the evolutionary features of linguliformeans appeared long before the Early Ordovician, including their burrowing adapta-

tion. After significant Late Cambrian loss of diversity, the generic diversity was restored during two pulses of Early Ordovician diversification, in the late Tremadocian and the late Early Ordovician (mid Arenig). By the Darriwilian, however, the linguliformeans had declined to subordinate elements in most environments, for example, displaced to marginal (shallower or deeper basin) environments. The craniiformeans lived free or cemented/ encrusted on the substrate with no trace of a pedicle and remained a comparatively minor group. They first appeared in the late Early Ordovician (early Arenig). Then their main radiation commenced in the late Darriwilian and continued into the Late Ordovician, with maximum diversity through the late Caradoc and early Ashgill. This included especially the development of the large, thick-shelled trimerellides, initially centered in island arcs of Kazakhstan and eastern Australia. The rhynchonelliformeans were arguably the most highly successful and completely diversified benthic group to inhabit Ordovician marine ecosystems. The group proliferated from an initial 4 Cambrian to 19 superfamilies through Ordovician time. In addition, the majority of the morphological adaptations exhibited by brachiopods developed during Ordovician time. The Early Ordovician record included the surviving Cambrian groups and two new higher-level groups, but, except for the orthidines and pentamerides, diversification levels remained at comparatively low levels. However, during the Mid Ordovician another five higher-level groups were added, and diversification levels increased dramatically. The Mid to Late Ordovician genus-level trajectories of 10 higher-level groups shows a dramatic, two-step, exponential rise of diversity, with early Mid Ordovician and latest Darriwilian to early Caradoc pulses to the first major diversification peak during the mid Caradoc. A sharp late Caradoc decline followed, then another rise to a higher, sharper early Ashgill peak. The large decline of diversity in the late Ashgill is attributed to the end Ordovician mass extinction. Individual diversification histories of a number of the higher-level taxonomic groups are also depicted. In the Ordovician Radiation the brachiopods played a significant role in establishing new adaptive strategies, life modes, and feeding habits. The marked diversity increases were accompanied by a general trend

29

30

 .  to greater specialization and decreased niche sizes to utilize the existing ecospace more completely. With low food requirements, brachiopods probably had advantages over competitors. Morphological innovations that developed include (1) the adoption of planar shell profiles allowing a free-lying, recumbent lifestyle by the strophomenides; (2) the appearances of more robust ball-and-socket articulations (cyrtomatodont teeth) for the hinge that clearly benefited the rhynchonellides, atrypides, athyridides, and spiriferides; and (3) appearance of spiral calcified ribbons to support complex lophophores in atrypides and earliest spiriferides. The diversity records of polyplacophorans (chitons) and symmetrical univalves (the tryblidiids— formerly “monoplacophorans”—and the bellerophontids) are reviewed by Cherns, Rohr, and Fry´da (chapter 18). The polyplacophorans diversified mainly in the Early Ordovician (especially Tremadocian to early Arenig). Occurrences are mainly associated with carbonates of the low-latitude Laurentian margins. Otherwise, their Ordovician record is temporally and spatially restricted. The tryblidiids similarly have highest diversities in the Early Ordovician and then decline through Mid and Late Ordovician time. In contrast, the bellerophontids have a low Early to early Mid Ordovician diversity, gradually increasing to a peak in the late Caradoc, and then steadily decline to the end Ordovician. Genus-level gastropod diversity, as a whole, is shown by Fry´da and Rohr (chapter 19) rising steadily from a low at the beginning of the Ordovician to a maximum in the late Caradoc–early Ashgill. But when the diversity records of some higher-level groups are dissected, two very distinctly different patterns emerge. The Archaeogastropoda (Selenimorpha), Mimospirina, and Archaeogastropoda (Trochomorpha)— though the diversification of the last group commenced later—show a similar pattern of rising diversity change. This contrasts markedly with the diversity records of two other major groups, the Euomphalomorpha and Macluritoidea. Again, their initial diversifications do not coincide, but both groups exhibit a remarkably similar Mid to Late Ordovician record—of increase to a peak in the early Mid Ordovician, then a dramatic drop to a plateaulike diversity low through Darriwilian to Late Ordovician time. Overall the gastropods exhibit four peaks of origina-

tion (late Early Ordovician, early Mid Ordovician, early Caradoc, and early Ashgill) and two extinction peaks. The first extinction in the early Darriwilian is less intense (it mainly affects euomphalomorphs and macluritoids), and the second is the largest, involving loss of nearly half the genera across all groups, and coincides with the end Ordovician mass extinction. In chapter 20 Cope documents the contrasting diversification patterns of the bivalve and rostroconch mollusks. A set of range charts illustrates the genus-level diversity patterns, and a summary of the diversity measures is also depicted. The first major radiation of bivalves in the Early Ordovician was limited geographically to the Gondwanan margins, across a wide range of latitudes, and to Avalonia, involving significant family-level diversification and a diversity peak in the early Arenig. Bivalves are known from many more localities in the Mid Ordovician, but the diversity did not increase markedly until later in the epoch and remained largely confined to Gondwana and surrounding areas. Then, in the early Caradoc (TS.5a), the second (and greatest) wave of expansion commenced after the migration and early diversification of bivalves in the low-latitude carbonate platforms regions of Baltica and Laurentia. The rostroconch mollusks, on the other hand, are a much smaller, infaunal group that exhibits a relatively high level of genus-level diversity in the Tremadocian, declined in extra-Gondwanan areas during the early Mid Ordovician, and then became reduced to near extinction at the end Ordovician. The nautiloid cephalopods were roving, commonly large predators and scavengers that occupied nektic to nektobenthic habitats in the Ordovician oceans and participated in their greatest ever expansion during Ordovician time. From one order at the beginning of the Ordovician, the group diversified to nine orders by the early Late Ordovician. The progressive nature of this radiation is outlined by Frey, Beresi, Evans, King, and Percival (chapter 21). The initial radiation during the Tremadocian involved the diversification of ellesmerocerids and first appearances of two other groups (endocerids, tarphycerids). These latter were the next to diversify, producing the first major diversification peak in the late Early Ordovician, and again were accompanied by first appearances of new groups (orthocerids, actinocerids, dissidocerids). These latter groups became actively involved in the

Introduction next rise in diversity during the Mid Ordovician, when again a number of new groups appeared. The second major diversification peak was in the Darriwilian and, unlike earlier radiation events, involved a significant biogeographic differentiation of the faunas into equatorial carbonate platform and high-latitude, mixed clastic and carbonate settings. Some of the biogeographically differentiated genera, however, suffered early extinction in the latest Darriwilian. The third major radiation in the Late Ordovician included appearances of further new orders and families. The faunas were diverse and more cosmopolitan, especially in the carbonate platform facies. A peak of diversification occurred in the early Ashgill, followed by rapid decline and extinction of families and genera in the end Ordovician mass extinction. Malinky, Wilson, Holmer, and Lardeux (chapter 22) document a number of problematic tube-shelled organisms of uncertain affinities. They include groups with an earlier (Cambrian) record, such as hyoliths, coleoloids, sphenothallids, and bryoniids, and two others, the cornulitids and tentaculitids, that are first confirmed in the Late Ordovician. Only the largest group, the hyoliths, has a significant biodiversity record, though, compared with its Cambrian record, the diversities and abundances are lower. The Ordovician members of the group are provincially well differentiated and diversified into higher-latitude assemblages, as well as forming a low diversity component in more diverse shelly assemblages. The provincial differentiation includes higher-latitude “Mediterranean” and intermediate-latitude “Baltic” provinces. Genus-level diversity increased from Early to Mid Ordovician time but declined in the Late Ordovician as the hyoliths of the Baltic province disappeared, probably as a result of the drift of Baltica into low latitudes. Faunas at the species level are profoundly endemic. Hyolith species diversity is low in the Tremadocian but rises toward the Early to Mid Ordovician boundary. Subsequently, in the Darriwilian, the diversity increased dramatically owing to the biogeographic differentiation of the faunas. The diversity continues to a maximum in the Caradoc and then loss of half the species in the Ashgill. Hints, Eriksson, Högström, Kraft, and Lehnert discuss aspects of biodiversity of four unrelated wormlike and sclerite-bearing groups in chapter 23. The groups include the scolecodonts, machaeridians, palaeo-

scolecidans, and chaetognaths. The organic-walled polychaete jaw apparatuses of scolecodonts provide a distinctive Ordovician diversity record. In the Early to early Mid Ordovician they have a low diversity; they then increase sharply in the late Darriwilian, as the most common genera diversified with a sharp peak of originations and accompanying increase of species diversity and abundances. After this rapid diversification event there was comparatively little diversity change—a slow Late Ordovician rise to a “high” in the latest Caradoc to early Ashgill and then decline toward the Ordovician-Silurian boundary. Sclerites are probably better known from the Cambrian (from groups such as the tommotiids and halkieriids) than in the Ordovician, but nevertheless the sclerite-bearing Ordovician machaeridians are widely distributed, with a relatively continuous record: (1) confirmed sclerites in the Tremadocian; (2) indications from the record in the latest Early Ordovician (early to mid Arenig) that a first radiation of the group had occurred; (3) further widening of distribution and increase in the number of taxa in the Mid Ordovician; and (4) evidence that suggests a continuation of the general trend toward a substantially higher diversity in the Late Ordovician. The other groups comprise the palaeoscolecidans of probably annelid or priapulid affinities and the chaetognaths (or arrow worms). They include comparatively few genera, with mainly long ranges, from the Cambrian through Ordovician to the Silurian (or younger). No marked Ordovician diversification or turnover events are recorded. Only two new paleoscolecidan genera and a new chaetognath genus originate in the Ordovician. Trilobite diversity patterns are addressed by Adrian, Edgecombe, Fortey, Hammer, Laurie, McCormick, Owen, Waisfeld, Webby, Westrop, and Zhou (chapter 24) in a presentation divided into three parts: (1) a survey of global (and Laurentian) taxonomic genus-level diversity patterns; (2) regional reports of taxonomic genus-level (and some species-level) diversity patterns for Australasia, South America, Avalonia, Baltica, and South China; and (3) a review of patterns of biofacies differentiation. The global data are presented using a nine-unit subdivision of Ordovician time and recognition of distinctive EFs, represented by family-level clusters—the Ibex Fauna I, Ibex Fauna II, and Whiterock Fauna. Each shows

31

32

 .  markedly different diversity trajectories. The Ibex Fauna I diversity was high in the Tremadocian and then declined; Ibex Fauna II diversified in the late Early Ordovician to early Darriwilian (Arenig) and then rapidly declined in the latest Darriwilian (both faunas failing to survive the end Ordovician mass extinction). Expansion of the Whiterock Fauna commenced in the Arenig, but the radiation was most intensive from the Mid Ordovician to the early Caradoc diversity peak. A slow initial decline followed, then more rapid decline toward the end Ordovician, but with more than two-thirds of these families surviving into the Silurian. The major diversification is recognized across three of the four provincial realms, that is, it is developed in the low latitudes of Laurentia, the midlatitudes of Baltica, and the higher latitudes of Gondwana (North Africa) and peri-Gondwana (Avalonia, Perunica), but the timing of the comparable radiation event is difficult to establish in Australia (part of low-latitude East Gondwana) because of a less than complete Mid Ordovician to mid Caradoc record. In terms of the environmental patterns during the radiation in Laurentia, the Whiterock Fauna dominated only the deep subtidal and reef habitats initially, in the early Mid Ordovician, but by the Late Ordovician, elements of this fauna had spread both onshore and offshore to virtually dominate all habitats across the environmental spectrum. Regional surveys exhibit some significantly different patterns. In Avalonia the Whiterock Fauna became dominant earlier, in the late Early Ordovician (early Arenig) across a wider range of shelf to slope habitats than in most other regions. In South America elements of the Whiterock Fauna appeared mainly in the mid Darriwilian and early Caradoc radiation pulses. In South China the main increase in the Whiterock Fauna occurred in the early Mid Ordovician, but the radiation was relatively less intensive from the Darriwilian to a mid Caradoc diversification peak than that shown on the global (and Laurentian) scale. In Australia the Whiterock Fauna first appeared in the Darriwilian, but its expansion is not recorded until much later, during the mid to late Caradoc. Trilobite faunas are shown to be widely distributed and highly adapted to benthic and pelagic ecosystems. The group adopted a wider range of feeding and habitat adaptations in the Ordovician than at any

other time—from the more richly populated shallower epipelagic and deeper mesopelagic biotopes, to the widest possible array of benthic biofacies developed across shelf to slope habitats, and across latitudinal zones. The benthic adaptation included (1) the olenid biofacies that occupied oxygen-deficient waters, with taxa apparently adapted to symbiotic living with sulfur bacteria; (2) the blind “atheloptic” trilobites that invaded deeper, poorly lit sites; (3) the suspension feeding trilobites, such as trinucleids and raphiophorids, that were more common in quieterwater, outer-shelf sites; (4) the small to medium-sized deposit feeders, such as the hystricurines and proetids, that occupied shallow-water carbonates, muds, and silts; and (5) the predators and scavengers, with their attached hypostomes, large size (e.g., asaphids), and well-developed eyes (e.g., phacopids, encrinurids), that largely dominated shelf environments. In chapter 25, Braddy, Tollerton, Racheboeuf, and Schallreuter outline aspects of the diversity of three other arthropod groups (eurypterids, phyllocarids, and ostracodes). The eurypterids are a small group of vagile (swimming and walking), mainly large predatory arthropods (water scorpions) that in Ordovician time occupied marine and probably some marginal marine settings. They have a poor fossil record, mainly fragmentary, though a few well-articulated skeletons are known; only about 20 sites have confirmed or doubtful records of eurypterids worldwide. The occurrences are mainly Late Ordovician, suggesting that the main diversification occurred during the Caradoc and Ashgill, though eurypterid trackways have been recorded from deposits without body fossils from an as yet unverified Early Ordovician age in South Africa. Phyllocarid crustaceans are widely distributed in time and space, especially in deeper-water black shale facies where the highest diversities and abundances occur. However, the taxonomy of the group needs to be reassessed. Preliminary work in South America suggests that rapid diversification occurred in the late Early Ordovician (early Arenig) and again during the early Caradoc. The ostracodes are a much larger group—again, not evenly or adequately documented in all parts of the world. The small, bivalved benthic forms (swimmers and crawlers) lived in marine waters of normal salinities on shallow shelves and shelf-slope habitats across a wide range of latitudes. Best-known records

Introduction of ostracode faunas are preserved in the carbonate facies of Baltoscandia, and these to some extent mirror the worldwide diversity patterns. The region is recognized as occupying higher to intermediate latitudes in the Early to Mid Ordovician but moved into the tropics during the Late Ordovician (chapter 5). Confirmed Early Ordovician biodiversity records are poor—only one early Tremadocian species but more diverse faunas in the late Early Ordovician (five families). More rapid diversification followed in the Mid Ordovician, especially in the mid Darriwilian. Significant diversity maxima occur in the Late Ordovician, first in the early Caradoc and then, after loss of about one-third of the species owing to the great Kinnekulle K-bentonite ash fall, higher peaks in the late Caradoc and mid Ashgill, prior to the dramatic decline to end Ordovician extinction of all species and loss of about a third of the families. In higher latitudes of Bohemia (Prague Basin), the ostracodes were preserved in clastics. No diversity record exists in the Early Ordovician, but the early Mid Ordovician has a low diversity; then there are three successive diversification pulses—in the late Darriwilian, mid Caradoc, and late Caradoc—before the Ashgill decline. Typically echinoderms occur in interbedded thin limestones and shales of shallow cratonic seas or locally in reefs, hardgrounds, or, less commonly, deltaic, slope, and deeper-water deposits across a wide range of latitudes. Echinoderm diversity of the five subphyla is fully described by Sprinkle and Guensburg (chapter 26). The class-level diversity more than doubled from levels in the Cambrian to the highest ever, class-level diversification of echinoderms by the early Late Ordovician. The diversity change of the five subphyla—that is, the four subphyla of Cambrian survivors (crinozoans, blastozoans, echinozoans, and homalozoans) and the new subphylum (asterozoans)— and the class- to order-level clades is depicted in a series of spindle diagrams. New clades were progressively added, from the 21 clades at the beginning of the Ordovician to the 30 clades in the Caradoc. Only a few groups exhibit significant species-level diversification during the Early to Mid Ordovician. The most remarkable diversification event for the echinoderms is during the mid Caradoc—(TS.5b–c), when the majority of groups, at almost all hierarchical levels (four out of five subphyla and all classes, genera, and

species), radiated dramatically over a relatively short interval of time. This diversification extended, as well, to an almost complete range of lifestyle modes. In terms of the two major Ordovician echinoderm groups, (1) the blastozoans were more common before the mid Caradoc expansion; (2) the crinoids achieved comparatively larger expansion during the mid Caradoc event; (3) both groups attained approximately equal importance in the latter part of the Ordovician; and (4) the crinoids became the dominant group after the latest Ordovician (Hirnantian), then remaining the larger group for the next 200 m.y. There was only a comparatively small decline in the numbers of higher-level groups affected by the end Ordovician mass extinction, given that, of the 27 groups present in Ashgill, 23 survived into the Silurian. Outlines of the lifestyle modes and the biogeographic relationships are also presented. The graptolites are a major planktic group with a worldwide distribution in sequences mainly representing outer-shelf, slope, and ocean basin habitats and geographic regions that range from low to high paleolatitudes. The species-level diversity analysis of the graptolites presented by Cooper, Maletz, Taylor, and Zalasiewicz (chapter 27) is based on a data set of species lists and zonal ranges from deeper shales of more or less continuous sections in three main regions: low paleolatitudes of Australasia, high to intermediate paleolatitudes of Avalonia (southern Britain), and the mainly intermediate paleolatitudes of Baltica. A set of standardized measures of diversity and turnover is employed in compiling the data. Patterns of diversity change, evolutionary rates, and environmental event significance are assessed. The initial diversification of these planktic organisms at the beginning of the Ordovician was slow, but by the late Early Ordovician (early Arenig) they had expanded spectacularly, especially in low-latitude Australasia, in a first of several successive waves of diversification, apparently each associated with global transgression and sea level highstand. Later waves of diversification were less intensive, the second occurring in the mid Darriwilian and a third in the mid Caradoc. Extinction events, in contrast, are recorded in the early Mid Ordovician (late Arenig), late Caradoc, and possibly late Darriwilian, and these are associated with regressive events, though they did not occur in all three regions simultaneously. In Australasia the graptolites came close to

33

34

 .  extinction in the late Ashgill, whereas in middle-high latitude Avalonia the decline started much earlier, during the mid Caradoc. The near extinction of graptolites in the late Ashgill is related to major regression and the end Ordovician (Hirnantian) glaciation. Paris, Achab, Asselin, Chen, Grahn, Nõlvak, Obut, Samuelsson, Sennikov, Vecoli, Verniers, Wang, and Winchester-Seeto analyze the global and regional species-level diversity patterns for chitinozoans in chapter 28. These organic-walled, flask-shaped microfossils are found only in Ordovician to Devonian sequences. The preferred view is that they represent egg capsules of an unknown, exclusively marine softbodied animal. The microfossils appear to have a pelagic distribution and are widely distributed in shallower to deeper environments as well as across paleolatitudes. The group includes 35 genera, which diversified more rapidly through the Early and Mid Ordovician than in the Late Ordovician. Only a few genera became extinct in the Early and Mid Ordovician, in marked contrast to the larger number (19 genera) that suffered extinction at the end of the Ordovician. The global chitinozoan species-level database shows a record of continuous diversification from the early Tremadocian to a first major peak in the late Darriwilian, then a short-lived interval of lowered diversity in the early Caradoc, followed by a rise to a second major diversity maximum in the mid Caradoc. For chitinozoans the initial major Early Ordovician (Tremadocian) radiation is the most significant (first-order) origination event. The most extreme (first-order) extinction event is at the end of the Ordovician, though there are less significant extinction pulses in the late Early Ordovician (early Arenig) and late Caradoc. Species diversity data for North Gondwana, Laurentia, Baltica, Avalonia, and South China were also analyzed to assess the nature of diversity change, turnover, and origination/extinction rates on a region-byregion basis employing rigorous sampling and counting methods. These results show some important regional differences. Albanesi and Bergström (chapter 29) discuss the diversity patterns of Lower and Middle Ordovician conodonts based on an extensive global database. These phosphatic microfossil elements are presumed to be from apparatuses in the cephalic part of a living eel-shaped pelagic animal that may have been dis-

tantly related to chordates. The elements are found in a wide range of marine environments, including the two broadly differentiated major biogeographic “Midcontinent” or “Atlantic” realms. A genus- and specieslevel biodiversity analysis of the Lower to Middle Ordovician conodonts is presented using diversity measures that display changing patterns and turnover rates. Overall, a progressive increase in diversity is shown for genera and species through six biostratigraphically distinct “intervals” of Early to Mid Ordovician time. New species arise at a rate of from 7.0 to 9.4 species appearances per m.y. Extinctions continue at a high rate with values of between 7.7 and 8.0 disappearances per m.y., though in one late Tremadocian “interval” the rate dropped to 2.5 per m.y. Assessments of holdover and carryover percentages were also calculated for genera and species. The results were comparatively high (more than 40 percent) for both holdover and carryover values, suggesting rapid diversity change throughout the entire Early to Mid Ordovician interval. Two significant diversification pulses are recognizable from the overall results, the first spanning from the late Tremadocian to near the end of the Early Ordovician and the second from the early Mid Ordovician to the mid Darriwilian. The Ordovician agnathan (jawless fish) and gnathostome (jawed fish) vertebrate diversity record is outlined by Turner, Blieck, and Nowlan (chapter 30). Two temporally and geographically distinct assemblages are recognized: in the Mid Ordovician of Gondwana, and in the Late Ordovician of Laurentia, Baltica, and Siberia. Only a few problematic taxa are recorded from the Early Ordovician (Tremadocian to late Early Ordovician), and all have doubtful affinities except for the confirmed small “bony” plate impressions of genus Porophoraspis from localities of early Arenig age in Australia. The more complete diversification of this Gondwanan vertebrate assemblage occurred in the Darriwilian of Australia and in the Darriwilian to early Caradoc interval of South America (Bolivia, Argentina). The greatest diversification, however, occurred in the mid Caradoc. Initially this was recorded only in Laurentia, with the dramatic appearance of at least 12 different types of vertebrate taxa in the Harding Sandstone and correlatives across North America. Taxa such as agnathans (including the first thelodonts) and the first stem-group gnathostomes were involved, and these components spread also to Baltica and

Introduction Siberia. Most elements of these Caradoc faunas were short lived. Then in the latest Ashgill (Hirnantian), during the interval of the end Ordovician Gondwanan glaciation, a further diversification of new vertebrate taxa occurred, apparently only in Baltica and Siberia. The timing of this event possibly coincides with the climatic amelioration immediately following the glaciation (chapter 9). In chapter 31, Nitecki, Webby, Spjeldnaes, and Zhen outline the species diversity patterns of two unrelated Paleozoic groups, the problematic receptaculitids and the algal cyclocrinitids, and review selected algal groups belonging to Chuvashov and Riding’s (1984) “Ordovician Flora” in order to clarify the timing of their initial diversifications. The receptaculitids were widely distributed in near-equatorial platform carbonates including reefs, especially in the Early to early Mid Ordovician. Their affinities are uncertain but probably linked to animals. They have a relatively continuous, low diversity record, but with weakly discernible peaks in the late Tremadocian to late Early Ordovician (early Arenig) and Mid Ordovician intervals and a higher peak within the late Caradoc. The cyclocrinitids had a wide distribution in carbonate and siliciclastic facies and are probably related to dascyclad algae. Their diversification commenced in the late Darriwilian, increased to a sharp diversity peak in the mid Caradoc, and then progressively lost species diversity to the end Ordovician. Like the cyclocrinitids, most of the other main Ordovician components of calcified green and red algae (e.g., dascyclads, udotaceans/codiaceans, solenoporans) exhibit diversity patterns with the greatest diversity change across the Mid to Late Ordovician boundary—initial appearances in the Darriwilian and significant diversification of these groups extending well into the Caradoc. In terms of timing, these diversification patterns in Ordovician calcified algae match in a rather similar way the major diversifications of a number of animal groups (e.g., bryozoans, corals), and they also played a part in the great late Darriwilian to Caradoc expansion of metazoan- and algal-dominated reefs (Webby 2002). The uncalcified brown algae also diversify significantly through the Late Ordovician (chiefly Caradoc). In contrast, the “blue-green algae,” which are mainly Cambrian cyanobacterial components, show a major decline in their importance through Ordovician time.

In chapter 32, Servais, Li, Stricanne, Vecoli, and Wicander review the acritarchs—their definition, biologic affinities, history of previous study, distribution, record of available literature, and data sets— and provide some new species-level data for South China and North Africa. This large group of organicwalled microfossils formed an important phytoplankton component of Ordovician oceans. The group has a markedly higher species diversity than in the preceding Cambrian. A global acritarch species database was compiled but remains difficult to use in establishing diversity curves because of the many doubtful taxonomic assignments and uncertain stratigraphic ranges reliant on earlier documentation. Similar problems exist in using the regional data sets (e.g., Baltica, North America), and in some sections the sampling remained inadequate. Species-level diversity data are therefore presented from only two regions—South China and North Africa. In South China, the diversity trend derived from eight biostratigraphically well-contrained acritarch assemblages shows a progressive increase from early Tremadocian to an early Mid Ordovician peak, then a slight decline during the early to mid Darriwilian. The North African record is mainly derived from borehole data, but significant stratigraphic breaks limit the continuity of the record to three stratigraphic levels: (1) a continuous Late Cambrian to mid Tremadocian record with diversity peak in the early Tremadocian; (2) a significant diversification record from the early Mid Ordovician (mid Arenig) to early Caradoc with the maximum diversity in the late Darriwilian; and (3) a shorter-ranging latest Ordovician diversity pulse, with first appearances of “Silurian”-type floral elements. These floral elements presumably reflect the climatic amelioration following the Ordovician glaciation (chapter 9), especially given the proximity of these occurrences to the south paleopole. Steemans and Wellman (chapter 33) record the progressive increase in the diversity of miospore (mainly cryptospore) taxa, from the Darriwilian through the Late Ordovician to a maximum across the Ordovican/ Silurian boundary. The end Ordovician climatic and glacial events that caused the mass extinctions in other fossil groups had no apparent adverse effects on the spore diversity. The abundant and cosmopolitan mid Darriwilian to Late Ordovician cryptospore assemblages are suggested to indicate the presence of a widely

35

36

 .  dispersed flora of small, nonvascular, bryophytelike plants on land, at least in the areas where damp conditions prevailed. In addition, an isolated occurrence of trilete spores in the latest Ordovician points to the initial development of vascular plants, possibly the antecedents to Siluro-Devonian rhyniophytes. This event may be linked with the postglacial amelioration of climate immediately following the glaciation (chapter 9). Evidence of actual Ordovician land plant megafossils, however, remains elusive. Mángano and Droser (chapter 34) review the trace fossil (ichnofossil) record in terms of their development, colonization trends, and ecologic diversity patterns. The development of macroboring activity and aspects of trace fossil utility in biostratigraphy are also outlined. A steady rise of ichnofossil diversity is recognized through Ordovician time, with the highest ichnodiversity developed during the Ashgill. The rise reflects the changes toward more diverse and complex behavioral patterns as new areas of ecospace became occupied and the development of more complicated deep tiering and pre- and postdepositional structures. The changing ichnodiversity patterns are assessed across a wide range of infaunal habitats in shallow marine (both siliciclastic and carbonate), deep marine, marginal marine, continental, and volcanic settings. During Ordovician time the shallow siliciclastic marine habitats developed more varied behavioral patterns, especially those formed by trilobites, and more distinctive tiering patterns, whereas in the shallow carbonate settings markedly increased bioturbation and tiering developed as a consequence of greater infaunal ecospace utilization. The deep marine environments also exhibit trends toward higher diversity and complexity, with the patterns changing from mainly feeding traces in the earlier Ordovician to more complicated, patterned grazing activity later in the Ordovician. In volcanic areas the diversity remained low, and the assemblages were of limited complexity, while in marginal marine settings there is evidence to suggest that some trilobites entered areas with brackish waters. Continental environments exhibit trackways that suggest that arthropods occupied coastal dunes, perhaps even early in the Ordovician. Miller (chapter 35) presents the final overview of the Ordovician Radiation, drawing on aspects of contributions in this volume and other relevant research published since the late 1960s. He notes that compi-

lations in the volume include some significant advances in the documentation of global diversity trajectories across a nearly complete suite of Ordovician taxonomic groups. The contributions provide new data from previously undersampled areas, a better understanding of phylogenetic relationships in some clades, and an improved global timescale—“a small sample of what has been—or could be—accomplished by the integration of new kinds of data and analyses . . . of Ordovician diversity.” The late Jack J. Sepkoski Jr. initially raised our awareness of the magnitude of this dramatic, extended radiation in his seminal papers of the 1970s, 1980s, and 1990s of compilations and analyses of global Phanerozoic marine diversity and showed how pivotal the Ordovician Radiation was in shaping the future of marine diversification worldwide. Miller’s chapter focuses mainly on two fundamental questions about the major radiation: why this “extensive, worldwide increase in total biodiversity” occurred during Ordovician time, and why “certain taxa radiated more appreciably than others.” These matters are discussed, using examples from chapters in the volume, in the context of (1) whether a global agent was principally responsible for shaping the diversity trajectory or whether local and regional factors also played an important part and (2) whether environmental and geographic limitations were overriding influences as well. Miller’s remarks conclude with five important key recommendations for future work. Briefly stated they are to (1) standardize analytical procedures; (2) dissect diversity patterns at finer geographic scales; (3) involve definitive integration of physical and chemical data with the biodiversity data; (4) undertake fieldbased studies across important sections to focus on biotic transitions; and (5) further investigate relationships between taxonomic diversification and morphological differentiation (disparity). ■

Future Directions

A great many things remain to be done in the future to better understand the nature, and likely causes, of the great Ordovician biodiversification event. Miller (chapter 35) has made suggestions to a similar effect, recommending a number of topics that need to be highlighted. Work initially should be directed toward completing, in a fully integrated and compre-

Introduction hensive way, the analyses that were started here: of taxonomic diversity patterns for all groups in time and space and across all taxonomic hierarchies from class to species. Other more exhaustive studies of ecologic diversity and morphological disparity should also be given priority. Then there are a host of other topics of relevance to Ordovician biodiversity studies that remain poorly understood, such as (1) the plankton ecology and productivity of the Ordovician oceans; (2) the overall impact of microbial life in all environmental settings of the Ordovician; (3) the reconstruction of trophic webs and assessment of nutrient levels; (4) fully integrated analyses of all the main Ordovician ecosystems, from terrestrial to the open ocean, and across latitudes; (5) fullest possible documentation of biodiversity data in at least two other major continental platform areas with complete and well-preserved Ordovician biodiversity records (e.g., Baltoscandia, South China) to rigorously test against the known patterns of onshore-offshore diversity change recognized across the North American platform; and (6) intensive documentation of biotas from the comparatively few well-exposed and well-preserved “windows” of island, island-arc, and oceanic segments. In a wider geologic context, we also need a much greater focus on the following topics: (1) global syntheses of Ordovician volcanic and orogenic histories; (2) a fully integrated and rigorously tested global Ordovician sea level curve; and (3) an improved understanding of global continental and oceanic configurations through Ordovician time—the recently published, cooperative quantitative biogeographic, paleomagnetic, and plate tectonic modeling approach (Lees et al. 2002) is a promising start. The only possible way to establish the real global significance of volcanic, orogenic, and plate tectonic effects on Ordovician

biodiversity is to have such fully comprehensive integrated worldwide analyses of these Ordovician volcanic, orogenic, and plate tectonic histories available to assess their respective roles. Finally, it is essential that more cooperative approaches are developed with other geologists, geochemists, and geophysicists, especially the scientists actively involved in past climatic and oceanographic modeling work. Future programs of Ordovician biodiversity studies should be more closely linked to the multidisciplinary groups working on earth-system processes and modeling of such past climatic and oceanographic change. The interactive roles of and responses to changing atmospheric compositions, ocean chemistry, and patterns of ocean circulation would have been critically important for the Ordovician biosphere and therefore are vitally important and relevant matters for our future understanding of the Ordovician radiations. As the ocean and climate states become better understood using the earth-system approach (see chapters 7 and 8), the extrinsic factors that may have been influential in shaping the major diversification events may be expected to begin to emerge more clearly.

 I thank Alan W. Owen, Arnold. I. Miller, and my editorial colleagues, Florentin Paris and Ian G. Percival, for their comments on the manuscript. Florentin Paris assisted also in the compilation of generic and specific data from authors, in the preparation of table 1.1, and provided advice on patterns of microplankton productivity. The chapter, like the many others in this volume, is a contribution to IGCP 410, “The Great Ordovician Biodiversification Event.”

37

part i

Scaling of Ordovician Time and Measures for Assessing Biodiversity Change

2

Stratigraphic Framework and Time Slices Barry D.Webby, Roger A. Cooper, Stig M. Bergström, and Florentin Paris

I

n compiling data for this volume, it was regarded as crucial for all participants to have access to a well-integrated Ordovician timescale based on the global and regional subdivisions, the key graptolite, conodont, and chitinozoan zonations (those with the greatest utility for wide-ranging correlation, though they are provincially distinct), and a set of closespaced global time lines. This provides the highestresolution correlation for the Ordovician biodiversity data collected and for global analysis of the data (for example, in determining the patterns of biologic and other significant marker events). In establishing a precise stratigraphic framework, it was vitally important to utilize as many cross ties between the provincially distinct zonations as possible and to integrate all the available, reliable radiometric dates to achieve the most finely calibrated scale possible. The main calibration involved a computer-based optimization approach (CONOP9 software) and is outlined by Sadler and Cooper in the following chapter. They compiled first and last appearances of more than 1,100 (mainly graptolite) taxa from almost 200 Ordovician and Silurian stratigraphic sections worldwide. This resulted in an ordered and scaled sequence of 2,306 biostratigraphic events within which stages and zone boundaries can be located. Twenty-two radiometrically dated samples were included in this sequence and serve to test the linearity of the scale and to calibrate it. Sadler and Cooper’s calibration results

in a significantly more highly resolved timescale than previously was possible and allows a particularly refined subdivision of 19 time slices to be employed. The time-slice intervals (TS ) represent durations of between 1.6 and 3.0 million years (m.y.) (on average ∼2.5 m.y.). With this level of refinement participants have been able to undertake comprehensive assessments of biodiversity change through Ordovician time, surveying a variety of different patterns of diversity, such as taxonomic richness, rates of speciation, turnover, species-genus ratios, and extinction rates in specific clades. These results allow more precise (temporally constrained) comparisons between the diversification patterns of different clade groups, and at different levels of the taxonomic hierarchy, than were previously possible. Based on the Sadler and Cooper calibration of the following chapter, the Ordovician is taken to range from 489 to 443 m.y., that is, it represents a duration of 46 m.y. ■

The Ordovician System and Its Systemic Boundaries

Lapworth’s (1879:11) original proposal and naming of the “Ordovician System” were influenced greatly by his recognition that the intermediate division of Lower Paleozoic rocks contained a fauna of “grand 41

Oncograptus m.-divergens maximus

victoriae lunatus

lunatus primulus protobifidus

fruticosus

fruticosus 3/4br fruticosus 4br frutico/approx

akzharensis

approximatus

approximatus

elegans

fasciculatus austro- sinicus dentatus zhejiang -ensis

HARJU

5c

Bu

foliaceus [= multidens]

Oa Ke Jo Ha

foliaceus

5b 5a

Id

Au

gracilis

Ku

gracilis

Ll

teretiusculus

Uh

teretiusculus

murchisoni

Ls As

distichus elegans

Ab

artus hirundo Fe

clavus caduceus imitatus

deflexus

clingani

Rk

VIRU

CHIENTANGK'G

ASHGILL CARADOC

clingani

ellesae

suecicus

5d

gibberulus victoriae Wh

simulans

varicosus fruticosus Mo

approximatus

?

Al

fasciculatus

Vl

lentus

Hn

3b

Sa

3a

hirundo

elongatus densus

2c

balticus

2b

phyllograptoides

2a

Bi

phyllograptoides

Ad. hunneb'gensis Triogr/Anisogr R. flabelliformis - parabola

Psigraptus R. flabelliformis - dichotomus

Ar. murrayi

Ar. murrayi

Mi

?

Vr

Ad. tenellus Cr

R. flabelliformis s.l.

4b

Vä

H. copiosus

Adelogr./ Clonogr.

4c

4a

Ln

Hu

Ad. victoriae Ad. victoriae / Pa. antiquus Psigraptus

jiangxi -ensis

bifidus

fruticosus 3br

Ar. pulchellus / Ar. macgillivrayi

NEICHIANSHANIAN

austrodentatus

murchisoni

Ch

KUNDA

dentatus

bicornis

linearis Na

St

LLANVIRN

callotheca

wilsoni

6a

Vo

gracilis

teretiusculus

morsus/upsilon m.-divergens maximus victoriae

R. scitulum/Anisogr.

GSSP 489 Ma

bicornis

complanatus

linearis

spiniferus clingani

Pi

VOLKHOV

austrodentatus

americanus

6b

OELAND

intersitus

spiniferus

pacificus anceps complexus Ca

(T S)

6c

Po

Ra

Pu

ARENIG

decoratus

pygmaeus

BALTOSCANDIA

persculptus extraordinarius

complanatus

complanatus

pygmaeus

ruedemanni

Hi

TREMADOCIAN

DARRIWILIAN

riddellensis

complexus

quadrimucronatus /johnstrupi

gracilis

gracilis

BRITAIN

persculptus extraordinarius/ ojsuensis mirus pacificus typicus

manitoulinensis

DARRIWILIAN

calcaratus

complanatus

DAWANIAN

spiniferus lanceolatus

MOHAWKIAN

kirki

WHITEROCKIAN

EASTONIAN GISBORNIAN

gravis

ornatus

YUSHANIAN

"pre- pacificus"

CINCINNATIAN

pacificus

IBEXIAN

BOLINDIAN

pacificus

ICHANGIAN

LOWER

GSSP

ORDOVICIAN

472 Ma

persculptus extraordinarius

TIME SLICE

CHINA

NORTH AMERICA

persculptus extraordinarius

uncinatus

BENDIGO CHEW CASTL'M YA

GSSP

MIDDLE

ORDOVICIAN

GSSP 460.5 Ma

AUSTRALASIA

LANCEFIELDIAN

GLOBAL SERIES

UPPER

ORDOVICIAN

GSSP 443 Ma

GRAPTOLITES

K. supremus

1d 1c 1b

Ad. hunnebergensis Pa

R. socialis/flabelliformis /desmograptoides /parabola

1a

FIGURE 2.1. Ordovician stratigraphic chart showing the correlations between the main graptolite zonal sequences of Australasia, North America, China, Britain, and Baltoscandia, the global Series, the global Tremadocian and Darriwilian Stages (see boxes surrounded by double lines), the mainly regional series and stages, and the 19 numbered and lettered time slices (TS ) (last column); at the left margin, the million-year divisions, specific radiometric ages (Ma), and IUGS-ratified global boundary stratotype sections and points (GSSPs) are shown. The formalized zonal graptolite taxa are cited here mainly using key species (or subspecies) only, except for the Tremadocian graptolite taxa, which are shown as customary with both generic and species names. Explanation of the abbreviated generic names: Ad. = Adelograptus, Ar. = Araneograptus, H. = Hunnegraptus, K. = Kiaerograptus, Pa. = Paradelograptus, and R. = Rhabdinopora.

Stratigraphic Framework and Time Slices distinctiveness” from older (Cambrian) and younger (Silurian) faunas. However, his proposal failed to win widespread early support because the Lower Paleozoic rocks in the British Isles had long been subdivided into two systems—Sedgwick’s Cambrian and Murchison’s Silurian—and the adherents of these opposed schools of thought preferred to maintain the bipartite division while continuing to argue about in which of the two systems Lapworth’s Ordovician rocks belonged. However, as the influence of Sedgwick and Murchison waned toward the end of the nineteenth century, Lapworth’s simple and practical compromise to use a tripartite subdivision of Lower Palaeozoic rocks gradually became more generally accepted and used by geologists worldwide. The name Ordovician was eventually adopted for global use by the International Geological Congress at Copenhagen in 1960 (Holland 1976). In terms of the two systemic boundaries, the first is the Cambrian/Ordovician boundary, and it is de-

FIGURE 2.1. (Continued ) Key to abbreviated regional names: Australasia: Chew = Chewtonian, Castl’m = Castlemainian, Ya = Yapeenian; China: Chientangk’g = Chientangkiangian; Britain: Cr = Cressagian, Mi = Migneintian, Mo = Moridunian, Wh = Whitlandian, Fe = Fennian, Ab = Abereiddian, Ll = Llandeilian, Au = Aurelucian, Bu = Burrellian, Ch = Cheneyan, St = Streffordian, Pu = Pusgillian, Ca = Cautleyan, Ra = Rawtheyan, Hi = Hirnantian; Baltoscandia: Pa = Pakerort, Vr = Varangu, Hu = Hunneberg, Bi = Billingen, Sa = Saka*, Vä = Vääna*, Ln = Langevoja*, Hn = Hunderum*, Vl = Valaste*, Al = Aluoja*, As = Aseri, Ls = Lasnamägi, Uh = Uhaku, Ku = Kukruse, Ha = Haljala, Id = Idavere*, Jo = Jõhvi*, Ke = Keila, Oa = Oandu, Rk = Rakvere, Na = Nabala, Vo = Vormsi, Pi = Pirgu, Po = Porkuni. The units shown with asterisks and compiled in a third, discontinuous Baltoscandian column are regarded as regional substages; the two other columns show regional series and stages. The former Baltoscandian Latorp stage name (not shown) is represented by its former lower and upper substages, now raised to full stage rank as the Hunneberg and Billingen stages (Meidla 1997). Note that in China a threefold series and a sixfold stage subdivision is now employed in relatively close conformity to international usages. However, some alternative stage names (and spellings) remain; for example, the Lower Ordovician Tremadocian has alternative names (either Ichangian or Xinchangian), and the succeeding Yushanian is alternatively the Dobaowanian. The Middle Ordovician includes most of the Dawanian and the Darriwilian (formerly Zhejiangian). The Upper Ordovician is subdivided into the Neichiashanian (alternatively, Aijiashanian), and the Chientangkiangian (alternatively, Qiangtangjiangian).

fined at the base of the conodont-based Iapetognathus fluctivagus Zone at Green Point, western Newfoundland (Cooper et al. 2001). This level is also the base of the lowest Ordovician global Tremadocian Stage (figures 2.1, 2.2). The second is the Ordovician/ Silurian boundary, and it is defined at the base of the graptolite-based Akidograptus acuminatus Zone (sensu lato) at Dob’s Linn, Scotland (Cocks 1988). This is one zone higher than the Glyptograptus persculptus Zone (figure 2.1). ■

Status of Global Series and Stages, and the Regional Subdivisions

Since 1974 the International Subcommission on Ordovician Stratigraphy (ISOS), under the aegis of the International Union of Geological Sciences (IUGS) and affiliated International Commission on Stratigraphy (ICS), has been actively focusing on how best to establish a single set of global series and stage divisions. In order to achieve these objectives it has been necessary to (1) identify significant levels (biohorizons) with wide-ranging global correlation potential using graptolites or conodonts, or both; (2) document best available sections as a basis for selecting a single global boundary stratotype section and point (GSSP) for global definition and correlation purposes; and (3) assign appropriate formal names for the chronostratigraphic intervals (Series and Stages) between the GSSPs (Webby 1995). The task has been a difficult one because of the highly provincial and ecologically differentiated Ordovician biotas (Jaanusson 1979), leading to a paucity of good stratigraphic sections containing associated graptolites, conodonts, and chitinozoans and the uneven distribution of reliable radiometric dates. In 1995, a 90 percent majority of voting members of the ISOS agreed on the use of a tripartite global (Series) subdivision of the Ordovician System into Lower, Middle, and Upper Ordovician Series. It was further accepted that each of the series would be further subdivided into two, giving a sixfold global Stage division for the Ordovician as a whole. A level near the base of the North American Whiterockian series was approved for defining the base of the Middle Ordovician (see figures 2.1, 2.2), and the base of the Nemagraptus gracilis Zone (now equated with the base of the British Caradoc) has been accepted (Bergström

43

WHITEROCKIAN

polonicus

Rg

472 Ma

variablis

anserinus

friendsvillensis

GSSP

inequal. kielcen.

suecicus

jenkinsi turgita subcylindrica ??

variabilis

sinuosa

norrlandicus

altifrons

originalis navis

flabellum/ laevis

triangularis

489 Ma

proteus

elong.-delt. gracilis IBEXIAN

ORDOVICIAN LOWER

GSSP

elegans

communis

deltatus/ costatus

Ha

??

langei C. brevis

Jo cervicornis Id dalbyensis

esthonica / raymondi

stentor

dianae

Sk

manitouensis angulatus

fluctivagus

merga

6b

nigerica

6a

robusta ? tanvillensis

Uh

tuberculata

Ls striata As

clavaherculi sebyensis

Al

5d 5c

?? dalbyensis

ponceti pissotensis clavata

5b 5a 4c

armoricana jenkinsi

formosa calix

4b

protocalix bulla henryi

4a

ornensis Vä

3b

Sa

3a

regnelli

Vl Hn Ln

cucumis

E. brevis

2c

baculata

2b

symmetrica

2a

Bi primitiva

symmetrica

1d

Hu Tripodus

?? low diversity interval

6c

deunffi

rhenana

amoenus St

ASHGILL

curvata

evae

andinus

Tl

multiplex

Ku

pirum

Bl

GSSP

Ancryochitina sp. 1

Ke

L. sp.A

serra

holodentata

Rk

hirsuta

??

reticulifera augusta

Oa

(T S)

oulebsiri elongata

fistulosa

Na

SLICE

barbata

barbata

CARADOC

gerdae

bergstroemi

LLANVIRN

compressa quadridactylus aculeata

tvaerensis

Tu

460.5 Ma

MIDDLE

alobatus

undatus

pygmaea/cristata spongiosa cancellata gracqui multispinata/duplicitas primitiva
S. sp. A


KUNDA

tenuis

sweeti

DARRIWILIAN

superbus

rugata

VOLKHOV

robustus velicuspis

gamachiana

Pi

Vo

hyalophrys/C. sp. 2

confluens Ch

vaurealensis senta

grandis Ma

anticostiensis

scabra/ taugourdeaui

ARENIG

Ri

GSSP

ORDOVICIAN

ordovicicus

HARJU

divergens

Po

gamachiana crickmayi

VIRU

CINCINNATIAN

shatzeri

Ed

MOHAWKIAN

UPPER

ORDOVICIAN

Ga

ellisbayensis/taugourdeaui

TIME NORTH GONDWANA

BALTOSCANDIA

NORTH AMERICA

OELAND

443 Ma

CHITINOZOANS

NORTH ATLANTIC

deltifer

Vr

angulatus

Pa

destombesi

TREMADOCIAN

GLOBAL SERIES

GSSP

CONODONTS NORTH AMERICAN MIDCONTINENT

conifundus

1c destombesi

1b 1a

FIGURE 2.2. Ordovician stratigraphic chart illustrating correlations between the main conodont zonal sequences of the North American Midcontinent and North Atlantic provinces, the chitinozoan zonal sequences of North America, Baltoscandia, and North Gondwana, the global Series, the Tremadocian and Darriwilian Stages, the mainly regional series and stages, the 19 numbered and lettered time slices (TS ) (last column), and the million-year divisions (at left margin). The formalized zonal conodont and chitinozoan taxa are cited here mainly using key species. In the chitinozoan sequences a few zonal intervals are defined informally using abbreviated generic name and informal (numbered or lettered) species, for example, C. sp. 2 = Conochitina sp. 2, L. sp. A = Lagenochitina sp. A, and S. sp. A = Spinachitina sp. A. In addition, two key taxa have been assigned the same species name; these are distinguished by having different generic names, namely, C. brevis = Conochitina brevis, and E. brevis = Eremochitina brevis.

Stratigraphic Framework and Time Slices et al. 2000) and ratified as the base of the Upper Ordovician (figure 2.1). The Sadler and Cooper calibration presented in the following chapter indicates that the global Series have the following durations: Lower Ordovician (17 m.y.), Middle Ordovician (11.5 m.y.), and Upper Ordovician (17.5 m.y.). Two global Stages have been ratified. The lowest division, the Tremadocian Stage, was adopted recently (Cooper et al. 2001), and the Darriwilian Stage, which is the upper division of the Middle Ordovician, was ratified earlier (Mitchell et al. 1997). The other four global Stage divisions remain unnamed, although GSSPs for the second and fifth Stages have recently been formally approved by the ICS and ratified by the IUGS (figures 2.1, 2.2). British regional series units may continue to be used by some Ordovician stratigraphers in a kind of de facto global nomenclature, but these names will be replaced gradually as the unnamed units become defined and ratified as global Stages. The formalized tripartite global Series (Epoch) usages of Lower (Early) Ordovician, Middle (Mid) Ordovician, and Upper (Late) Ordovician should now be maintained. These global Series usages differ from earlier, more traditional regional concepts of “lower,” “middle,” and “upper” as employed particularly in North America and Baltoscandia (as well as the former Soviet Union). In North America the Ibexian has been assigned to the lower Ordovician but now includes an interval of topmost Cambrian beds (equivalent to three conodont zones) in its lowermost part (Cooper et al. 2001). The Whiterockian and Mohawkian were combined as representing the middle Ordovician, and the Cincinnatian was retained as upper Ordovician (figure 2.1), but Sweet (1995) advocated that the Mohawkian and Cincinnatian be combined as “upper Ordovician,” which is more consistent with the scope of the global Upper Ordovician Series as recently ratified by the ICS.

FIGURE 2.2. (Continued ) Key to abbreviated regional names: North American Midcontinent: Sk = Skullrockian, St = Stairsian, Tl = Tulean, Bl = Blackhillsian, Rg = Rangerian, Tu = Turinian, Ch = Chatfieldian, Ed = Edenian, Ma = Maysvillian, Ri = Richmondian, Ga = Gamachian. For key to other abbreviated names and other symbols, see figure 2.1.

In the British Isles five regional series are recognized, instead of six as previously, with the “Llandeilo” now abandoned in favor of incorporating the lower part in an expanded Llanvirn, and including the upper part in the overlying Caradoc (Fortey et al. 1995, 2000). The “Llandeilo” encompassed the interval of the graptolite-based Husteograptus teretiusculus and Nemagraptus gracilis zones combined, consequently straddling the newly formalized MiddleUpper Ordovician boundary. Therefore the series usage of “Llandeilo” should be discontinued. A more restricted conception has been retained by Fortey et al. (1995, 2000), named the “Llandeilian stage” and comprising the upper part of the Llanvirn series (figure 2.1). In this volume we have drawn a clear distinction between the Llandeilian stage, and the former, expanded, series usage of “Llandeilo.” The latter term is quoted in parentheses in this volume because a few workers have continued to recognize the old name. In Baltoscandia a tripartite subdivision into the Oeland (lower Ordovician), Viru (middle Ordovician), and Harju (upper Ordovician) series has been used. The Oeland series includes the Pakerort to Kunda interval (figure 2.1). A similar tripartite series subdivision of lower, middle, and upper Ordovician was applied across wide tracts of Russia and parts of Asia (Sokolov et al. 1960). Further information about the regional series and stage subdivisions shown in figures 2.1 and 2.2 is presented in Webby (1998), Fortey et al. (2000), and Chen et al. (2001). ■

Toward a Biostratigraphically Integrated Zonal Framework

The existing global and regional stratigraphic frameworks with their accompanying successions of biozones that represent unidirectional, nonreversible evolutionary change comprise, for all practical purposes, chronostratigraphic scales (Harland et al. 1990). The biostratigraphically most useful schemes for wide-ranging correlation are the graptolite, conodont, and chitinozoan zonal indices, but long-distance correlations are complicated because of the marked biogeographical and ecological differentiation even among these groups of organisms. Consequently, each representative group exhibits at least two biogeographically distinct zonal successions (figures 2.1, 2.2).

45

46

 .   . Pacific-type provincial graptolites are present in the Australasian and North American columns, representing low paleolatitude regions, and European-type (“Atlantic”) provincial graptolite zones in the British columns of figure 2.1, representing high paleolatitudes (Skevington 1973; Berry 1979). On the other hand, the Baltoscandian and Chinese columns have mixed Pacific and European faunal affinity (Cooper et al. 1991). Cooper et al. (1991) found ecological depth stratification to have a marked effect on graptolite distribution, while Finney and Berry (1997) recognized graptolite abundance in Nevada to be concentrated along the continental margins. The conodont zonal successions include both the North Atlantic type and the North American Midcontinent type, based on data from Bergström (1986), Sweet (1995), and Sweet and Tolbert (1997). Sweet (1995) and Sweet and Tolbert (1997) additionally employed a graphic analysis approach for correlation of the Midcontinent province conodonts. The chitinozoan zonal successions also include distinctly different provincial assemblages from North America, Baltoscandia, and the higher paleolatitude regions of North Gondwana (Paris 1996). Bergström (1986) has presented the most comprehensive documentation of cross ties between the different representative zonal successions. His survey extended to the graptolite and conodont successions of northwestern Europe and North America. A total of 44 separate zonal ties were recognized between the European graptolite and North Atlantic conodont successions through the Ordovician of northwestern Europe (mainly in Britain and Baltoscandia), and a further 34 zonal ties were identified between Pacifictype graptolite and North Atlantic conodont successions across North America. These cross-zonal ties served to establish a much more firmly integrated biostratigraphically based correlation framework for the chronostratigraphic Ordovician timescale. The much revised timescale compilation illustrated in figures 2.1 and 2.2 adds many further cross-tie adjustments to establish the present correlation framework. It uses the same Bergström cross-tie approach, not only for the graptolites and conodont data but also for available data from the three most important, provincially distinct chitinozoan successsions (figure 2.2).

■

Definition of Time Slices (TS )

The six primary time slices, TS.1–6, more or less coincide with the six approved global Stages (figures 2.1, 2.2). Formalized Tremadocian and Darriwilian Stages coincide with TS.1 and TS.4, respectively; the other four Stages remain unnamed. The time slices are further subdivided into a total of 19 secondary divisions using the supplementary lowercase lettering a, b, c, and d. In most cases the boundaries dividing these smaller units are tied to important graptolite or conodont zonal boundaries, or both. The Tremadocian Stage is subdivided into four time slices (TS.1a–d), with TS.1a broadly equating with the Lower (Early) Tremadocian (and exhibiting the classic Rhabdinopora flabelliformis series), and TS.1b–d correlative with the Upper (Late) Tremadocian of Cooper (1999a). In terms of the Baltoscandian regional subdivisions, TS.1a equates with the Estonian Pakerort stage, TS.1b is correlative with the Estonian Varangu stage, and TS.1c–d is equivalent to the Swedish lower-middle Hunnebergian (Cooper and Lindholm 1990). The unnamed upper Lower Ordovician Stage is subdivided into TS.2a–c, which are broadly correlative with the lower-middle parts of the British Arenig. The TS.2a is equivalent to the wide-ranging, distinctive Tetragraptus approximatus Zone (and partly correlative T. phyllograptoides Zone of Baltoscandia), and TS.2b–c more or less span the Swedish Billingen and the Australasian Bendigonian to Chewtonian stages. The unnamed lower Middle Ordovician Stage (equivalent to the Australian Castlemainian and Yapeenian stages) is represented by TS.3a–b, and the base of TS.3a more or less equates with the bases of the North American Whiterockian series (the Rangerian stage), the Australian Castlemainian stage, and the Baltoscandian Volkhov stage. Fortey et al. (2000) commented on the difficulties of correlating the base of the Middle Ordovician into parts of northern Gondwana and associated terranes, where the British regional Arenig series subdivision is applied. The base is equated with a level in the mid Arenig (probably within the Whitlandian stage), but this correlation is regarded by Fortey et al. as speculative. In South Wales where the regional stages of the Arenig are defined (Fortey and Owens 1987), in par-

Stratigraphic Framework and Time Slices ticular, through the Whitlandian interval, the continuity of the faunal succession is incomplete, and there is a lack of diagnostic ties to the graptolite zonation recorded elsewhere. Consequently, in the higher paleolatitude regions of North Gondwana and associated terranes it may not always be easy to differentiate confidently between the biotas of TS.2 and TS.3. The North Gondwanan chitinozoan zonation presented in figure 2.2 (and the acritarch counterparts) may, however, provide more precise stratigraphic control for use in the clade group analyses of these biotas across North Gondwanan regions, and the differentiation between the upper Lower Ordovician (TS.2) and the lower Middle Ordovician (TS.3) is less ambiguous. The Darriwilian Stage is defined as including TS.4a–c. The TS.4a interval coincides with the Undulograptus austrodentatus Zone (and its Subzones in China; Chen and Bergström 1995), and the TS.4a–b boundary coincides with the base of the British Llanvirn series (figure 2.1). Consequently, the uppermost part of the British Arenig extends into the lowest part of the Middle Ordovician (TS.4a). The TS.4b–c boundary is equivalent to the Baltoscandian series boundary between the Oeland (top of Kunda stage) and the Viru series. The unnamed lower Upper Ordovician stage is represented by TS.5a–d and spans slightly more than the entire British Caradoc series. The TS.5a coincides with the remarkably widespread Nemagraptus gracilis Zone. The base of the North American Mohawkian series (and base of the Turinian stage) correlates with the base of TS.5b. The base of the Baltoscandian Harju series (and Nabala stage) coincides with the base of TS.5d. The TS.5b–c boundary has been equated with the Kinnekulle K-bentonite (volcanic ash bed) at the base of the Estonian Keila stage (Männil and Meidla 1994) in Baltoscandia, and its possible correlative, the Millbrig K-bentonite bed in North America, that defines the base of the Chatfieldian stage, near the

middle of the Mohawkian (figure 2.2). This supposedly correlative, trans-Iapetus ash marker has been dated at 454.6 Ma (Huff et al. 1992; Leslie and Bergström 1995). However, Min et al. (2001) have recently cast doubt on this intercontinental correlation, citing significant differences in the geochemical data of primary mineral phases and a wide discrepancy (about 7 m.y.) between their apparent 40Ar/39Ar ages, as evidence against the Kinnekulle and Millbrig K-bentonites being a part of a single eruptive event. Though only a few apparent radiometric ages have been determined based on U/Pb data (e.g., Tucker and McKerrow 1995), they may be more reliable and do not exhibit such markedly dissimilar apparent ages. The unnamed upper Upper Ordovician stage is represented by TS.6a–c and spans most of the British Ashgill series (except the lower part). The lower part of TS.6a is recognized by the wide-ranging Dicellograptus complanatus Zone. The TS.6b more or less coincides with the Paraorthograptus pacificus Zone. The TS.6c correlates with the British Hirnantian stage and is approximately equivalent to the North American Gamachian stage and the Estonian Porkuni stage. The British column shows the correlation of the mainly shelly faunal-based Ashgill stages (figure 2.1) with the standard graptolite zonation, following Fortey et al. (2000). In the Fortey et al. scheme the Caradoc-Ashgill boundary is placed in the upper half of the Pleurograptus linearis Zone (e.g., upper TS.5d). However, Rickards (2002) recently identified beds in the middle part of the Rawtheyan of the type Cautley area in northern England as having a P. linearis Zone age, which suggests that the accepted Fortey et al. correlation may need to be revised. The Hirnantian incorporates the end Ordovician glaciation in its lower two-thirds (chapter 9). There are two recognizable phases of extinction, the first at the base of the Hirnantian (base of TS.6c) and the second within the Glyptograptus persculptus Zone in the mid–late Hirnantian (mid–late TS.6c).

47

3

Calibration of the Ordovician Timescale Peter M. Sadler and Roger A. Cooper

O

rdovician deep-water shales contain the prerequisites for a high-resolution timescale: rich successions of graptolite faunas, datable ash-fall Kbentonites, and minimally interrupted accumulation. Traditionally, the first appearances of selected graptolite taxa define provincial sets of zones, into which radiometrically dated bentonites are subsequently arrayed to achieve a numerical timescale. Provincial differences and the modest numbers of zones impose the primary limits on resolution. We have taken a different approach that avoids the constraint of zones. ■

The New Approach

To achieve a unifying timescale for this book, we applied computer-assisted optimization to combine graptolite range charts from all provinces directly, without using zones. The optimization process searches for a model sequence of all range-end events that best fits all the locally observed taxon ranges. Dated ashfall events are included with the range-end events from the outset of the search; thus, they too receive optimal placements in the model sequence and permit numerical ages for biostratigraphic events to be estimated by interpolation. By this method, we interpolated ages for the Australasian graptolite zone boundaries (VandenBerg and Cooper 1992). Ages for the boundaries in other subdivision schemes were

48

derived, using all available criteria to correlate into the Australasian graptolite zonation (chapter 2). ■

The Raw Data

The timescale is based on an ordered and scaled sequence of 2,306 events: 22 dated bentonites that are associated with graptolites or other fossils for which contemporary graptolites are known (table 3.1); 12 undated bentonite beds that help tie together short sections from the Mohawk Valley (C. E. Mitchell pers. comm.); and the first and last appearances of 1,136 taxa (1,119 graptolites, plus 17 trilobites and conodonts) as reported from almost 200 stratigraphic range charts worldwide, from the basal Ordovician to early Devonian. The range charts represent Arctic (31), Cordilleran (10), Midwestern (6), and Northeastern (29) North America, as well as South America (9), Great Britain (24), Iberia (9), Germany (10), Scandinavia (16), eastern Europe (14), the Middle East to Central Asia (14), Siberia (1), China (16), and Australasia (9). These are all the range charts we could find that meet modern standards of graptolite systematics and depict collections from shaley facies. Inclusion of Silurian range charts and radiometric dates permits a more robust calibration. CambroOrdovician trilobite and conodont taxa were added to improve the constraints near the base of the Or-

Calibration of the Ordovician Timescale TABLE 3.1. Radiometric Control Points Age (Ma)

Unit

491 ± 1 [U-Pb] 489 ± 0.6 [U-Pb] 483 ± 1 [U-Pb] 469 +5 -3 [U-Pb] 465.7 ± 2.1 [U-Pb] 464 ± 2 [U-Pb] 464.6 ± 1.8 [U-Pb]

Latest Cambrian Cambrian/Ordovician Late Tremadocian Arenig/Llanvirn Llanvirn Llanvirn Llanvirn

460.4 ± 2.2 [U-Pb]

Llanvirn

455 ± 3

[Ar-Ar]

“Kinnekulle” K-bentonite

456.9 ± 1.8 [U-Pb]

“Kinnekulle” K-bentonite

454.8 ± 1.7 [U-Pb]

“Pont-y-ceunant ash”

457.4 ± 2.2 [U-Pb]

“Pont-y-ceunant ash”

453.1 ± 1.3 [U-Pb]

“Millbrig” K-bentonite

454.1 ± 2.1 [Ar-Ar]

“Millbrig” K-bentonite

454.5 ± 0.5 [U-Pb]

“Deicke” K-bentonite

445.7 ± 2.4 [U-Pb] 438.7 ± 2.1 [U-Pb] 436.2 ± 5.0 [Ar-Ar] 430.1 ± 2.4 [U-Pb] 423.7 ± 1.7 [Ar-Ar] 421.0 ± 2 [K-Ar] 417.6 ± 1.0 [U-Pb]

Ashgill Llandovery Llandovery Llandovery/Wenlock Ludlow Ludlow Lochkovian

Stratigraphic Constraint Peltura scarabaeoides scarabaeoides Peltura scarabaeoides scarabaeoides Hunnegraptus sp. Undulograptus austrodentatus above Didymograptus artus [Cerro Viejo section—Argentina] Holmograptus spinosus Didymograptus murchisoni Pterograptus elegans Hustedograptus teretiusculus Climacograptus bicornis Climacograptus wilsoni Climacograptus bicornis Climacograptus wilsoni Dicranograptus clingani Diplograptus foliaceus Dicranograptus clingani Diplograptus foliaceus Climacograptus bicornis Ensigraptus caudatus Climacograptus bicornis Ensigraptus caudatus Climacograptus bicornis Ensigraptus caudatus [Dobs Linn section—UK] [Dobs Linn section—UK] Coronograptus cyphus Oktavites spiralis Lobograptus scanicus Neodiversograptus nilssoni Monograptus uniformis

Reference Davidek et al. 1998 Landing et al. 2000 Landing et al. 1997 Tucker & McKerrow 1995 " Huff et al. 1997; Mitchell et al. 1998 Tucker & McKerrow 1995

(12) (13) (14)

"

(15)

"

(18)

"

(18)

"

(19)

"

(19)

"

(20a)

Kunk et al. 1985; Tucker & McKerrow 1995 (20a) Tucker & McKerrow 1995

(20b)

" " " " Kunk et al. 1985 Tucker & McKerrow 1995 Tucker et al. 1998

(21) (22) (23) (24) (25)

Note: The “stratigraphic constraint” column names the local range chart in which the dated bed can be placed or the taxon used to constrain an isolated bed in the optimal sequence. P. s. scarabaeoides is a trilobite; all other listed taxa are graptolites. “Sensitive High Resolution Ion Microprobe” (SHRIMP) dates (Compston and Williams 1992) were avoided pending resolution of systematic differences from the isotope dilution dates and questions about analytical standards. The entries “U-Pb” and “Ar-Ar” distinguish dates determined by the uranium-lead and argon-argon methods, respectively.

dovician, where graptolite species richness is very low. Because habitats and fossil preservation are patchy, the local range charts disagree in detail concerning the sequence of first and last appearances of taxa. The rules for resolving these discrepancies are straightforward, but the data set is so large that computer assistance is essential. ■

Computer-Assisted Calibration

Optimization algorithms in the CONOP9 software (Kemple et al. 1995; Sadler 2001) proceed by iterative improvement from an initially randomized sequence of events toward a “best-fit” sequence. They arrive at a model sequence for all 2,306 events with the best-known fit to the field observations in the sense that a minimum of stratigraphic range extensions is required to fit every local range chart to the model. The means of measuring the length of range extensions is critical to the outcome of the opti-

mization. For this calibration task, we minimized the number of other range ends overlapped by the extensions, not the stratigraphic thickness of the extensions as minimized in graphic correlation techniques. This change eliminates bias due to variations in accumulation rate and favors those sequences preserved in the most richly fossiliferous sections. Such a misfit measure can solve ordinary correlation problems in which all sections span approximately the same time interval. It needs augmentation here because the local sections span only small fractions of the total time interval. We ensure that local sections are “stacked” in the correct order by simultaneously minimizing a second component of misfit borrowed from unitary association techniques (Guex 1991; Alroy 1992). The additional term counts the number of coexistences of pairs of taxa that are implied by the model sequence but not observed anywhere. The outcome of searching on these two criteria is an optimally ordered sequence of events.

49

 .    .  Position in Scaled, Optimally Ordered Sequece 750 410

1,000

1,250

1,500

LOWER ORDOVICIAN 420

1

2

1,750

2,000

MIDDLE ORDOVICIAN

3

2,250

2,500

2,750

UPPER ORDOVICIAN

4

5

3,000

3,250

3,500

Tucker et al. 1998 T&M 25 Kunk et al. 1985

6

430

Millions of Years Before Present

50

T&M 24 T&M 23

440

T&M 22

Akidograptus ascensus FAD

6

T&M 21

Dicellograptus complanatus* FAD

450

5 460

T&M 18,19,20a,b Nemagraptus gracilis FAD

T&M 15

4 470

3

Undulograptus formosus* FAD Tetragraptus recl. reclinatus* FAD

T&M 13,14; Huff et al. 1997 T&M 12

Regression (R 2 = 0.99): y = 2 × 10−6 x2 − 0.038x + 527

2 480

T. phyllograptoides* FAD

1 490

Landing et al. 1997 Landing et al. 2000 Davidek et al. 1998

Vertical error bars: two standard deviations Horizontal error bars: range of placements within best-fit solutions

500

FIGURE 3.1. A projection of the six numbered chronostratigraphic time-slice boundaries, as used in this book, from the scaled optimally ordered sequence into the numerical timescale, by regression on the optimal placements of dated ash-fall events. We use, somewhat arbitrarily, a very small second-order polynomial term and force the line through the high-quality Cambro-Ordovician date. Similar results can be achieved with cubic splines (Frits Agterberg pers. comm.) or the locally estimated sum of squares. T&M refers to numbered dated items in Tucker and McKerrow (1995), many of them from Tucker et al. (1990).

Biostratigraphic field observations alone support the optimal sequence, but severe simplifying assumptions are unavoidable for scaling the time intervals between events. Two explicit simplifications were chosen as the least objectionable: in the long term, net taxonomic change (numbers of first- and lastappearance events) is a guide to the relative duration of whole sections; in the short term, stratigraphic thickness is a better guide to relative duration, especially in deep-water shales. The scaling assumptions are applied as follows. First, all local range charts are adjusted to fit the optimal sequence: observed taxon ranges are extended, if necessary, and missing taxa are inserted. Second, the total thickness of each section is rescaled according to the fraction of the composite sequence that it spans. Third, the spacing of adjacent events in the best-fit composite sequence is set equal to the average of all the local rescaled spacings. Zero

values are included in the average to allow for mass extinctions and rapid radiations. The result at this stage is a relative timescale—a scaled, optimally ordered sequence in which events are spaced in proportion to one possible approximation of their separation in time. ■

Tests of the Calibration

The ultimate test of the scaling process plots the position of the dated events in the putative relative timescale against their radiometric ages on a regular numeric timescale. The near-linear regression in figure 3.1 justifies the scaled, optimally ordered sequence as a plausible proxy for a timescale and therefore suitable for interpolating the age of traditional zone boundaries. This regression test was actually applied in three stages. The first stage omitted the Australasian composite sec-

Calibration of the Ordovician Timescale tion (VandenBerg and Cooper 1992) and used only the 607 taxa observed in more than one non-Australasian section; the second stage added Australasian ranges; the third stage (figure 3.1) added those taxa known from only a single section. The fit to a linear regression improved slightly with each step. Because of provincial differences, presumably aggravated by the difficulty of achieving perfect optimization for huge data sets, the ordinal regression between our optimally ordered sequence and the Australasian sequence is not perfect, but it has a high correlation coefficient. Australasian zone boundaries were placed in the spaced, optimally

ordered sequence at the first appearance of a cluster of zonal taxa (usually three to four). If this level did not coincide with the appearance of the customary defining species, a substitute species was selected from the cluster (asterisks in figure 3.1).  This work was supported in part by National Science Foundation grant EAR 9980372 to Sadler. Fritz Agterberg, Barry Webby, and Henry Williams suggested improvements on an earlier draft.

51

4

Measures of Diversity Roger A. Cooper

M

easuring diversity change through geologic time faces difficulties from several sources. The obvious ones include variable preservation quality and collection completeness. Incomplete collecting will tend to shorten the stratigraphic ranges of taxa and reduce diversity. It will also be unlikely to detect the rare species, affecting diversity estimates as well as origination and extinction rates. A less obvious bias stems from the nature of most biostratigraphic data sets and is discussed in this chapter. The basic data for the study of diversity change through geologic time are generally in the form of stratigraphic range charts. They are compiled from either individual sections or isolated collecting localities. When one is dealing with single sections, the stratigraphic ranges of taxa can be plotted in terms of the stratigraphic thickness of strata through which the taxon ranges. But when one is dealing with several sections, or with broad regions, the ranges of taxa are generally recorded in terms of biostratigraphic units such as zones or chronostratigraphic units such as stages. Fossil data collected from isolated localities must be assigned to stages or zones and compiled into composite zonal (or stadial) range charts. For large-scale, and global, studies such as the Great Ordovician Biodiversification Project (IGCP 410), ranges expressed in terms of zones, stages, or larger units are the norm. The problem of converting stratigraphic ranges of taxa into measures of diversity and taxonomic rates

52

of evolutionary change has been extensively discussed in the literature (Harper 1975, 1996; Sepkoski 1975; Raup 1985; Sepkoski and Koch 1996; Foote 1997a, 1997b, 2000a, 2000b with numerous references; Alroy 2000). A range of measures has been suggested that attempt to avoid introduced bias. However, it appears that no one measure of diversity, or of species origination or extinction intensity, avoids all problems (Jablonski 1995; Foote 2000a). The purpose of this chapter is to examine some of the more commonly used measures and assess which is the most appropriate for deriving diversity patterns from range charts. Commonly used measures of rates of origination, extinction, and faunal turnover are also outlined. ■

Mean Standing Diversity (MSD)

Ideally, the aim of paleontological diversity measures is to estimate the standing taxic diversity of a broad region through geological time (Sepkoski 1975; Sepkoski et al. 1981), equivalent to gamma diversity or total species richness (R, Rosenweig 1995). Standing taxic diversity is the total species diversity at a given instant in time. Because of the difficulty in directly measuring the standing diversity at a given instant in geologic time, we try instead to estimate the MSD over a specified time interval. This involves counting species present in strata that span the time interval. But there are several ways of expressing the number

Time

Measures of Diversity

c

Time unit (2 m.y.)

d b

a Total diversity (dtot) = 4 Species/m.y. (di) = 2 Normalized diversity = 2.5 (dnorm)

FIGURE 4.1. The four ways in which a species can be present in a time interval; a, range through; b, originate within the interval and range beyond it; c, range into the interval and terminate within it; d, confined to the time interval. The three measures of diversity for the time interval are shown.

of species present and of allowing for introduced biases. The different methods can produce very different diversity curves for the same data set. There are four ways in which a taxon can be present in a time interval, illustrated in figure 4.1: (a) range from the preceding interval to the following interval (range-through taxa); (b) originate within the interval and range into the following interval; (c) range from the preceding interval and become extinct within the interval; (d) confined to the time interval. For the purposes of the present exercise, range ends in categories b–d that coincide with an interval boundary are regarded as lying within the interval. ■

Effects of Time Interval Duration

A principal source of bias derives from unevenness in time interval duration (Sepkoski and Koch 1996; Foote 2000a). Sepkoski and Koch recommend minimizing this problem by combining zones of short duration and subdividing zones of long duration in order to minimize their differences. Foote (2000a) has modeled the effects on measures of diversity, and origination and extinction rates, of time interval duration, preservation quality, “edge effects,” and the taxonomic rates themselves. He concluded that single-interval taxa produce many undesirable dis-

tortions, and he recommended that diversity and rate measures be adopted that do not count singleinterval taxa (taxa confined to a single time interval). Rather than counting the taxa within a time interval, he recommended counting only those taxa whose stratigraphic ranges cross interval boundaries. However, this procedure would preclude a significant proportion of paleontological data for some fossil groups (chapter 27) and is unlikely to be acceptable to most paleontologists. Some of the concerns raised by Foote may be minimized by using a precise timescale with subequal time intervals, as recommended by Sepkoski and Koch (1996). As discussed later in this chapter, by using model data sets that are comparable to those found in the fossil record, it is possible to test directly the performance of a range of diversity measures in estimating MSD through time intervals that range widely in duration. ■

Estimators of MSD

Three commonly used estimators of MSD are total taxic diversity, taxa per million years (m.y.), and normalized total taxic diversity here referred to as normalized diversity. In the following discussion the species is taken as the taxonomic unit, but genera and other taxonomic categories can be substituted as appropriate.

Total Diversity (dtot ) The simplest and most commonly used measure for estimating MSD is total diversity (dtot )—the total number of species (or other taxa) that are recorded from the time interval.

Species per m.y. (di) This measure is simply total diversity divided by the duration of the time interval (i ) in m.y. and allows for the probability that longer time intervals will capture more species. This is a rate measure and is therefore not strictly comparable with the other two measures, which are counts. The absolute value yielded by the measure is less important than its trend through time.

53

54

 . 

Normalized Diversity (dnorm ) Species ranges in zonal range charts will overestimate true ranges because few species ranges will completely span the zones in which they first appear or last appear or to which they are confined. To compensate for this, Sepkoski (1975) devised a diversity measure here referred to as the normalized diversity measure. It is the sum of species that range from the interval below to the interval above, plus half the number of species that range beyond the time interval but originate or become extinct within it, plus half those that are confined to the time interval itself. The measure also normalizes for variability in time interval duration to the extent that the longer a time interval is, the more species will begin or end within it or are confined to it. ■

Model Data Sets

For the purposes of testing and comparing the three diversity measures, six model data sets have been

constructed. The durations of time intervals and of species life spans in the data sets are comparable with those found in the graptolite data sets for Australasia, Baltica, and Avalonia (chapter 27). In the model data sets, both time interval durations and species longevities are uneven, as in most paleontological data sets. For simplicity, it is assumed that species ranges are complete and that all species that can be present in a bed are present. In addition, it is assumed for the purposes of the exercise that the observed ranges are the true, global ranges. The six model data sets give a total of 40 time intervals. The smallest set contains five time intervals, which range from 1 to 2.5 m.y. in duration. This set is shown in figure 4.2, which also shows how the zonal range (thin lines), as recorded in range charts, considerably overestimates the true range (bold lines) for most taxa. Twenty-nine species are present and average 2.49 m.y. in duration. The largest set (set 6) contains 10 time intervals, which range in duration from 1.5 m.y. to 5 m.y. Eighty-three species are present and range in duration from 0.3 to 3 m.y., averag-

FIGURE 4.2. Model data set 1 comparing the three measures of diversity—total diversity (dtot ), species/m.y. (di ), and normalized diversity (dnorm )—with mean standing diversity (MSD; shaded area).

Measures of Diversity TABLE 4.1. Properties of the Six Trial Data Sets and Comparison of Trial Diversity Measures with Mean Standing Diversity (MSD) Data Sets

1

2

3

4

5

6

5 6 2 3.80 29 4.50 16 11 14 13

6 7 1 3.30 37 5.40 37 17 18 15

8 4 1 2.20 62 3.90 67 22 35 37

10 4 1 2.10 76 6.20 115 33 41 36

10 5 1.5 1.80 83 2.60 62 30 41 47

Properties of model data sets Number of time units Longest time unit Shortest time unit Mean duration of time units Number of species (total diversity) Mean duration of species Number of range through species Number of confined species Number of originations Number of extinctions

5 2.5 1 1.80 29 2.49 15 11 14 13

Net % difference from MSD (allows for +ve and −ve) Total diversity (dtot ) Species/m.y. (di ) Normalized diversity (dnorm )

13.1 −8.2 −5.9

17.1 −17.7 −2.0

20.0 −29.3 −4.0

36.8 −18.0 −9.2

47.5 −41.8 −10.5

39.5 31.9 −21.5

Mean 29.0 −13.9 −8.8

Mean % difference from MSD (regardless of +ve or −ve) Total diversity (dtot ) Species/m.y. (di ) Normalized diversity (dnorm )

15.3 17.3 7.4

23.6 26.4 9.0

20.3 29.7 5.6

28.6 20.1 9.8

29.1 29.9 14.1

30.1 52.8 13.0

Mean 24.5 29.4 9.8

ing 2.6 m.y. The basic properties of the data sets are given in table 4.1. Because the exact stratigraphic ranges of species in the model data sets are known, the values for MSD can be precisely calculated for each of the time intervals. In the first data set (figure 4.2) it is highest in time interval D, where it reaches 10.6. At any given time in interval D there are, on average, 10.6 species in existence. Time interval B has more species than interval C yet scores a lower MSD. This is because more of the species in interval B have ranges that do not span the entire interval. Generally, this level of precision is not available in the fossil record, and species are recorded only as present or absent in each time interval (shown as the zonal range in figure 4.2). The performance of the three measures in estimating MSD can thus be directly tested against its “true” value for each time interval in each data set (figure 4.2 and table 4.1). Two measures of comparison with MSD are given (table 4.1). The first is the mean net difference from MSD, expressed as the cumulative difference in value from MSD for each measure in each time interval. The lowest value indicates the closest comparison. However, positive and negative differences cancel each other out, so that it is possible for a curve to be the mirror image of MSD yet score a low value. The second measure of comparison with

MSD is the mean percentage difference from MSD, expressed as the cumulative difference, regardless of whether it is positive or negative. The second measure requires the two curves to have similar shapes and similar magnitude in order to score a low value and is most meaningful in the present context. ■

Results of Trials

In model data set 1 (figure 4.2) it can be seen that no one measure closely approximates true MSD in all time intervals. Total diversity (dtot ) overestimates MSD in all time intervals except interval C. In interval D it overestimates MSD by 80 percent. The species per m.y. (di ) curve generally underestimates MSD. The normalized diversity (dnorm ) measure also underestimates MSD but is the best estimator. All three measures indicate a diversity peak in interval B where none exists. The performance of the three measures through 40 time intervals in the six trial data sets is shown in table 4.1 and figure 4.3. The total diversity (dtot ) measure consistently and sometimes substantially overestimates true MSD, deviating from it by 25 percent on average, and the species per m.y. (di ) measure generally underestimates it, deviating by 29 percent on average. The species per m.y. measure is the most erratic of the three. Normalized diversity (dnorm ) is

55

 .  20

15 10

10

5

5

0

0 1

2

3

4

30

Set 3 6ⴛ37

15

Set 4 8ⴛ62

25

Species

Set 2 5ⴛ37 Species

Species

20

20 15 10

1

5

2

Time Units

3

4

5

6

5

Time Units

30

0

Set 5 10ⴛ76

25

Species

1

2

3

4

5

6

7

8

9

10

Time Units

20 60

15 10

Set 6 10ⴛ83

50

5 40

0 1

2

3

4

5

6

7

8

9

10

Time Units

Total diversity (d tot ) Species per m.y. (d i ) Normalized diversity (d norm )

Species

56

30 20 10 0

Mean standing diversity (MSD)

1

2

3

4

5

6

7

8

Time Units

FIGURE 4.3. Comparison of alternative diversity measures in model data sets 1–5. The normalized diversity measure (dnorm ) consistently gives the closest estimate.

consistently the closest approximation of true MSD, deviating by 10 percent on average. The MSD curve in all model data sets is a smoother curve than the species per m.y. curve, which inserts, or exaggerates, peaks and troughs, sometimes in mirror image of the MSD curve. The normalized diversity curve copies the shape of the MSD curve more faithfully than either total diversity or species per m.y. The difference between total diversity and MSD diminishes as the number of taxa that begin, end, or are confined to the time interval diminishes. This happens when larger taxonomic categories with long stratigraphic ranges (genera, families, orders) are used or when very fine time intervals are used. If all taxa present in a time interval range from the interval below to the interval above, then total diversity will equal both MSD and normalized diversity. Conversely, when low-level taxonomic categories such as species are used, a higher proportion of them will begin, end, or be confined to the time interval. Total diversity becomes a poor estimator of MSD in these circumstances. Using time intervals that are relatively long in comparison with the ranges of the taxa produces similar results. Two further diversity measures were also tested. The first (Sepkoski et al. 1981) uses the normalized

diversity measure discussed earlier but further normalizes for the difference between the mean duration of a species in the data set and the mean duration of a time interval. The second diversity measure was a composite derived from all four measures, the most appropriate being chosen for each time interval as judged by a range of criteria. However, in neither case was there a significant improvement on the normalized diversity curve in estimating MSD. ■

Measures of Evolutionary Change

Apart from the effects of regional migration, diversity within a region is a function of origination (o), here taken to mean a cladogenetic event, and extinction (e). The combination of origination and extinction within any time interval gives the faunal turnover (o + e). Similarly, net increase or decrease in diversity is given by originations minus extinctions (o – e). The raw data for these four parameters of evolutionary change will be influenced by two main factors. The first is total diversity (d ) in the time interval; high-diversity time intervals are more likely to have higher counts for each measure. The second factor is duration of the time interval (i ); longer time

Measures of Diversity intervals are likely to have higher counts for each measure. For these reasons Sepkoski and Koch (1996) recommend that a variety of measures be used for these evolutionary parameters. For extinctions they recommend (1) the total number of extinction events in the time interval (the “raw count”), (2) the percentage of extinction, which is the number of extinction events divided by the total number of taxa present, expressed as a percentage, and (3) the mean per capita extinction rate, which is the percentage of extinction divided by the duration of the time interval in m.y. Following are the three measures along with their equivalents for origination, faunal turnover, and net diversity increase/decrease. Number of extinctions Percentage of extinction Per capita rate of extinction

=e = ed × 100 = edi

Number of originations Percentage of origination Per capita rate of origination

=o = od × 100 = odi

Faunal turnover Percentage of faunal turnover Per capita rate of faunal turnover (in m.y.)

= (o + e) = (o + e)2d × 100 = (o + e)2di

Net increase/decrease Percentage of net increase/ decrease Per capita rate of net increase/ decrease

= (o – e) = (o – e)2d × 100 = (o – e)2di

■

Conclusions

The conversion of stratigraphic range data to diversity estimates involves inescapable biases. None of the three measures tested faithfully estimates MSD for data sets in which both species longevities and time interval durations are nonuniform, that is, for most paleontological data sets. Of the measures tested, that most commonly used—total diversity count (dtot )—generally overestimates MSD. Species per m.y. (di ) produces diversity curves that fluctuate more strongly than MSD and can have inflections that mirror those of MSD. It is the most erratic of the three measures tested. The normalized diversity measure (dnorm ) consistently gives the closest estimate of true MSD and is recommended for use in studies of diversity change through geologic time. Because origination, extinction, faunal turnover, and net diversity increase/decrease in a time interval are influenced by the interval’s duration and total diversity, the measures that normalize for these biases (percentages and per capita rates) should be given in addition to the raw counts.  I thank Drs. Arnie Miller and James Crampton for their comments. The project was supported in part by the Marsden Fund, Royal Society of New Zealand.

57

part ii

Conspectus of the Ordovician World

5

Major Terranes in the Ordovician L. Robin M. Cocks and Trond H. Torsvik

F

rom Jurassic times onward, magnetic stripes from spreading centers are preserved on the modern ocean floors, and unraveling these are good guides to the migration of terranes through time. However, the old ocean floors present in the Ordovician have been either lost through subduction or distorted and displaced through obduction, and thus the identification and positioning of the terranes present at that time must rely on less direct methods. Chief among these are paleomagnetism, but that can indicate only paleolatitude, not paleolongitude, and faunal studies, which can indicate terrane separation and to a general extent paleolatitude, but in a more subjective way than paleomagnetism. In addition, the distribution of distinctive facies can be helpful for some periods, particularly, for example, at the time of the latest Ordovician glacial event. We have summarized our methods elsewhere (Cocks and Torsvik 2002) and collaborate to place the Ordovician terranes in positions that satisfy the current constraints from paleomagnetism, faunas, and sediments. The faunal evidence to support the reconstructions is reviewed in Fortey and Cocks (2003). However, the degree of confidence varies greatly with both place and time. In general, the present-day North Atlantic area, which in the Ordovician included Laurentia, Baltica, and West Gondwana and their adjacent terranes and island arcs, is well constrained. However, the many terranes that today make up most of Asia are less

securely defined and their Ordovician locations much less certain. In Ordovician times there was one supercontinent, Gondwana, occupying paleolatitudes from the South Pole to north of the equator, and three other major terranes, Laurentia (largely North America), Baltica (northern Europe), and Siberia. There were also many other smaller but still substantial terranes, mostly surrounding Gondwana and known as peri-Gondwanan; chief among them were Avalonia, Sibumasu (ShanThai), Annamia (Indochina), South China, and North China. Smaller but faunally distinctive terranes included Perunica (Bohemia), the Taurides and Pontides of Turkey, Tarim, the Kazakh terranes, the Himalayan terranes, and the Argentine Precordillera. In addition to these there were several island arcs in the various oceans, many of which carried faunas that have survived. These terranes are reviewed in turn, with appropriate reference to their changing positions as the Ordovician progressed. In addition to the terranes discussed in this chapter, Apulia and the Hellenic Terrane (both in southern Europe) and various Mexican terranes are shown on our maps, but since there is little evidence for their existence and detailed positions in the Ordovician, they are not discussed further here. In this short summary the various major terranes are listed and their positions presented in four maps (figures 5.1 to 5.4) depicting Early Ordovician (480 Ma, 61

62

.  .    .  FIGURE 5.1. Southern Hemisphere Ordovician terrane disposition in latest Tremadocian and base Arenig (about 480 Ma) time (from Cocks and Torsvik 2002: figure 4). 13a, Apulia; 13b, Hellenic; 14a, Taurides; 14b, Pontides; 15a, Lhasa; 15b, Qiantang; Kar., Karakum; Kaz, Kazakh terranes (shown arbitrarily). The three small areas with asterisks between Laurentia and Baltica denote the only three places on the Iapetus island arcs from which reliable paleomagnetic data are known. North China is absent because it was entirely within the Northern Hemisphere. The thick dashed and dotted line surrounds core Gondwana. Figures 5.1 to 5.3 are Schmidt’s Equal Area projection, with projection center at the South Pole.

late Tremadocian and earliest Arenig), earliest Late Ordovician (about 460 Ma, early Caradoc), and latest Ordovician–earliest Silurian (approximately 440 Ma, Ashgill-Llandovery) times, respectively. The first three figures show only the Southern Hemisphere, where most of the Ordovician terranes lay, but the fourth shows the whole earth. The latter emphasizes how much of the Northern Hemisphere was occupied by the vast Panthalassic Ocean. ■

Gondwana

Core Gondwana consisted of a very large area, divided for descriptive reasons into two, West and East Gondwana. West Gondwana included Africa, Madagascar, Florida, South America, and Arabia. East Gondwana included Greater India, Antarctica, New Guinea, and most of Australia. Core Gondwana is shown by a distinctive dashed line in our figures. Although Armorica (including Iberia) is shown outside that line as a peri-Gondwanan terrane, it is now thought that it did not split off from West Gondwana until well after the end of the Ordovician. This

is because (a) the Ordovician faunas of Brittany, the Montagne Noire, and the Iberian Peninsula are essentially the same as those of North Africa (Morocco, Algeria, and Libya) and (b) the position of Gondwana is now better constrained by both paleomagnetic and faunal data than it was until recently, all of which negates published models showing a substantial seaway between Armorica and Gondwana. In the earliest Ordovician, West Gondwana also included Avalonia and Perunica, but they became separate from it during the period (see later in this chapter). Arabia has been thought by some authors to have been formed from several terranes (termed Lut, Sanand, and others), but there seems little evidence that it formed anything other than part of core Gondwana in Ordovician times. Gondwana formed at about 550 Ma (Meert and Van der Voo 1997), and its early dispersal history commenced with the rifting off of Avalonia from the northern South America–northwestern Africa part of West Gondwana in Arenig time. The whole continent occupied more than 100 degrees of paleolatitude in the Ordovician and thus spanned the area from the

Major Terranes in the Ordovician FIGURE 5.2. Southern Hemisphere Ordovician terrane disposition in early Caradoc (about 460 Ma) time (base map derived from Cocks and Torsvik 2002: figure 5). The thick dashed line surrounds core Gondwana, and some putative spreading centers and subduction zones are shown. The terrane names are shown on figure 5.1. North China was entirely within the Northern Hemisphere.

South Pole to north of the equator. In the Early Ordovician the South Pole lay under North Africa, perhaps Libya, but as time and drift proceeded, by the end of the Ordovician the pole lay under central West Africa or perhaps slightly farther to the west under adjacent Brazil (Cocks and Fortey 1988). ■

Baltica and Kara

Baltica follows the traditional outline as used by many authors (references in Cocks and Fortey 1998) but includes the Malopolska and Lysogory areas of the Holy Cross Mountains, Poland (Cocks 2002), and the hidden southward extension of the Urals as far south as the northern Caspian Sea (Cocks 2000: figure 6). It excludes the highest nappes in the Trondheim area, Norway (which were part of Laurentia), and also the Taimyr region of Siberia, which had previously been considered as part of Baltica (Cocks and Fortey 1998). North Taimyr is now considered to be part of the Kara Terrane and central and southern Taimyr as parts of the main Siberian continent. We have restored Novaya Zemlya to its pre-Triassic posi-

tion as a direct extension of the Urals (Torsvik and Andersen 2002). The whole terrane rotated by more than 100 degrees between the Vendian and the Middle Ordovician, and in particular 55 degrees of that rotation took place in the Late Cambrian and Early Ordovician (references in Torsvik and Rehnström 2001), and thus today’s southwestern Tornquist margin in central Europe faced toward Laurentia in the Early Ordovician. Baltica moved from fairly high temperate paleolatitudes in the earliest Ordovician northward to more equable climates by the Late Ordovician. The movement of the Kara Terrane, which includes northern Taimyr and Severnaya Zemlya, shown in figures 5.1 to 5.4, is well constrained paleomagnetically (Torsvik and Rehnström 2001); the terrane includes Ordovician faunas, but these are not yet well evaluated. ■

Siberia

Although persisting with the traditional use of this name for an Ordovician terrane its margins were very

63

64

.  .    .  different from those of Siberia today. We follow the outline of Rundqvist and Mitrofanov (1993), which includes most of the Baikalides, which accreted on to the main Angaran craton in the center of Siberia in the Late Precambrian. We also include south and central Taimyr as an integral part of the terrane. The paleomagnetic data are good (Smethurst et al. 1998), and from them we can see that the paleocontinent moved from south to north across the paleoequator and into northern low to temperate latitudes during the Ordovician, making it, North China, and parts of Laurentia and East Gondwana (Australia and New Zealand) the only lands of consequence yet identified from the then Northern Hemisphere. More important, Siberia underwent rotation after the Ordovician, so that in our maps today’s north faces southward. ■

Laurentia

Ordovician Laurentia included most of today’s North America north of the Ouachita Front, in addition to Greenland, Spitzbergen, and all the British Isles north of the Iapetus suture (following the fit of Bullard et al. 1965). It also included the eastern part of today’s Siberia, the Chukhot Peninsula. From both paleomagnetic and faunal evidence Laurentia appears to have occupied a transequatorial position with little movement during the whole Ordovician, and we have no evidence of any substantial rotation during the period.

times, and drifted northward across the narrowing Iapetus Ocean, leaving a widening Rheic Ocean to its south. Avalonia docked, probably softly, with Baltica at about 443 Ma, the close of the Ordovician. The movement of the terrane is well constrained paleomagnetically (Torsvik et al. 1993). ■

The Sibumasu, sometimes termed the Shan-Thai, Terrane stretches today from Burma (Myanmawr) in the north, through western Thailand and western Malaysia, to Sumatra in the south. The paleomagnetic data are scanty, but the faunas place the terrane close to South China in the peri-Gondwanan collage (Fortey and Cocks 1998). ■

Annamia

This terrane occupies nearly all of Indochina. There are few paleomagnetic data, and the faunal descriptions date from the early twentieth century or before. Nevertheless, enough is known to understand that it had little in common with Sibumasu, with which it did not unite until the late Jurassic, but that it too was also a member of the peri-Gondwanan collage in the Ordovician. The Early Ordovician faunas reviewed by Fortey and Cocks (2003) suggest the relatively high paleolatitude shown in figure 5.1. ■

■

Sibumasu

South China

Avalonia

This region (Cocks et al. 1997) includes the eastern seaboard of North America north from Cape Cod, Massachusetts, through the Maritime Provinces of Canada, eastern Newfoundland, southeastern Ireland, England, Wales, Belgium, and Holland, and eastward to the Trans-European Suture Zone. There is no convincing evidence of it being split between “East” and “West” Avalonia in the Ordovician—the use of these terms should be discontinued. To the south Avalonia is bordered today by the various terranes of Gondwana and peri-Gondwana, notably Armorica and Perunica (Bohemia). In the earliest Ordovician Avalonia formed an integral part of West Gondwana but separated from it, probably in Arenig

The margins shown in figures 5.1 to 5.4 follow the outline of Rong et al. (1995), and we show the modern coastline only to aid recognition in the diagrams. There are some paleomagnetic data for the Ordovician, and the terrane appears to have moved from temperate to tropical paleolatitudes during the course of the period. There is much faunal evidence (discussed in Fortey and Cocks 2003) to place the terrane near East Gondwana. ■

North China

This terrane, which also includes the Korean Peninsula, also follows the outlines of Rong et al. (1995). From both the fauna and the paleomagnetic data we

Major Terranes in the Ordovician FIGURE 5.3. Southern Hemisphere Ordovician terrane disposition in latest Ordovician– earliest Silurian (about 440 Ma) time (base map derived from Cocks and Torsvik 2002: figure 6). Siberia and Tarim are absent because they were entirely within the Northern Hemisphere. Part of North China (with today’s outline of the Korean peninsula) is to be seen at the top right. The other terrane names are shown on figure 5.1. The thick dashed line surrounds core Gondwana, and some putative spreading centers and subduction zones are shown.

can see that it was not too far from Laurentia in the Early Ordovician but moved from north to south across the paleoequator, and by the latest Ordovician it appears to have approached, or even formed part of, the peri-Gondwanan terrane collage. It eventually docked with South China in the Permian.

■

Perunica

The Bohemian Massif of central Europe, which forms the core of the old terrane of Perunica, was an integral part of West Gondwana until some time in the Early Ordovician. However, certainly by early

FIGURE 5.4. Global reconstruction for latest Ordovician–earliest Silurian time (about 440 Ma), from Cocks and Torsvik (2002: figure 7). Mollweide projection, showing all the major areas absent from figure 5.3 and the extensive Panthalassic Ocean.

65

66

.  .    .  Caradoc times (figure 5.2) it was adrift in the Rheic Ocean on both faunal (Havlícˇek et al. 1994) and paleomagnetic (Tait et al. 1997) grounds. It remained as a separate small terrane until Early Devonian time and during the period appears to have undergone some rotation. ■

Taurides and Pontides

The Pontides, which are developed north of the east-west Anatolian Fault, and the Taurides, which are developed to the south of that fault and north of the main Arabian plate, form the greater part of modern Turkey. There are few paleomagnetic constraints, but the faunas (Dean et al. 2000) indicate separation of the two terranes, with the Pontides at a higher latitude off West Gondwana than the Taurides in the Ordovician. Early Ordovician brachiopods from the Taurides (Cocks 2000) confirm this intermediate paleolatitude. ■

Tarim and the Kazakh Terranes

The many terranes that make up central Asia between the Urals (to the west of which was Baltica) and the Siberian craton are the least well constrained in figures 5.1 to 5.4: little reliable paleomagnetism is known from them. S¸engör and Natal’in (1996) have identified more than 20 terranes in the area, which they postulated as forming an enormous Kipchak Arc between Baltica and Siberia in the Lower Paleozoic, which subsequently collapsed and coalesced to form the central Asian collage of today. Many of these structural units have Precambrian cores, and a large number of differing Ordovician benthic and planktonic faunas are known from this large area. Nikitin et al. (1991) show the modern extent of some of these units. However, recent analysis of the faunas (Fortey and Cocks 2003) has demonstrated that at least some of these areas—for example, the Chingiz, Tien Shan, and Chu-Ili terranes, as well as Tarim—have affinities with Gondwana and peri-Gondwana rather than with Baltica or Siberia. Thus these areas are shown only diagramatically as two small units in figures 5.1 to 5.4, apart from Tarim, which is shown near South China—this is based on a combination of

faunal similarity and some paleomagnetic data from the Devonian. ■

The Himalayan Terranes

Figures 5.1 to 5.4 show the Afghan Terrane, the Lhasa Terrane, South Tibet (15a in figure 5.1), and the Qiangtang Terrane, north Tibet (15b in figure 5.1), and their boundaries are well delineated from Upper Paleozoic and later faunal and tectonic studies. There are Ordovician localities, many with faunas, scattered at intervals through these Himalayan regions. However, there are no reliable Lower Paleozoic paleomagnetic data from these areas, and the faunal data have not yet been intensively evaluated for the Ordovician; thus we show these terranes as tied to Gondwana in our reconstructions. ■

Precordillera

The Precordillera terrane, which today lies in an area in the northwest of Argentina, carries Cambrian and Early Ordovician faunas of Laurentian affinity. Whether the Precordillera actually formed an integral part of Laurentia or lay close offshore as a separate terrane is uncertain. However, it clearly became separate from Laurentia soon afterward, since faunal endemicity there was at its highest in Caradoc times (Benedetto 1998). The terrane docked with Gondwana sometime during the Silurian, sandwiching the Famatina island arc between it and the supercontinent. ■

Island Arcs

Figures 5.1 to 5.4 omit island arcs, apart from three small areas in the Iapetus Ocean representing paleomagnetic data from intra-arc localities in eastern North America published by Mac Niocaill et al. (1997). However, there were at least two island arcs in the Iapetus Ocean in the Ordovician (Mac Niocaill et al. 1997) that amalgamated with each other and with the neighboring continents as the Iapetus narrowed during the Ordovician and closed in the Silurian. There were further island arcs in the Chukhot area to the north of Laurentia (Natal’in et al. 1999), the Famatina range between the Precordillera and

Major Terranes in the Ordovician Gondwana in South America (Benedetto 1998), and off Gondwana in what is today southeastern Australia (Percival and Webby 1996), from all of which Ordovician rocks and various faunas are known (Cocks 2001). In addition, the substantial area preserved

within central Asia today (see Kazakh terranes and Tarim earlier in this chapter) certainly carried islandarc faunas in the Ordovican, but the identity, extent, disposition, and tectonic evolution of the various arcs there are not yet certain.

67

6

Isotopic Signatures Graham A. Shields and Ján Veizer

H

ere we review current understanding of the isotope chemistry of Ordovician seawater as determined from published analyses of the isotopic compositions (strontium, neodymium, sulfur, carbon, and oxygen) of marine authigenic precipitates. Isotopic data have implications for our understanding of global dynamics and paleoenvironmental evolution; however, many published interpretations are still controversial. In general, the past 20 years have witnessed a laudable departure from using bulk samples to carefully selected calcite components, such as microstructurally pristine brachiopods and early marine cements, as subjects of geochemical study.

■

Strontium Isotopes

Because of the long, 2–5 m.y., residence time of strontium in the oceans, the 87Sr/86Sr ratio of seawater is the same worldwide and can be used for both global stratigraphic correlation and interpretation of major trends in planetary dynamics (Veizer 1989). Seawater 87Sr/86Sr ratios decrease worldwide across the Cambrian-Ordovician boundary from >0.7091 to ∼30‰ to 1>1province - Ashgill) Province(late (late Darriwilian Darriwilian–Ashgill)

1

0.4 4 provinces

fene fene

3 provinces

0.3

ptilo

rhab

2 provinces

0.75

amplex ptilo

0.2

0.5

hallop amplex hallop

0.1

0.25

cyst

cyst esthon cyclo

0 4b 4c 5a 5b 5c 5d 6a 6b 6c Time Slice time-slice

0

North American

cyclo

Baltic

Siberian

Mediterranean

Province province

FIGURE 16.6. Geographic aspects of Ordovician bryozoan radiation. Frequency histogram on the left shows varying levels of provinciality in terms of the proportion of genera recorded in two or more of bryozoan provinces through Ordovician TS.4b to 6c (see figures 16.1, 16.3, 16.4). Differences in the relative proportions of genera belonging to eight major stenolaemate clades in these four provinces are depicted in the diagram on the right.

153

154

   .                  The number of endemic genera in each province is 19 for North America, 41 for the Baltic, 7 for Siberia, and 1 for the Mediterranean. Changing patterns of provinciality are evident when the numbers of genera present in two, three, or all four provinces are quantified through time (figure 16.6). The largest spread of genera between provinces occurred in TS.5d (latest Caradoc–earliest Ashgill), the only interval for which a genus (Hallopora) has been recorded from all four provinces. Pending a more sophisticated analysis of these distributional data, it appears that bryozoan provinciality was greatest in Arenig times and declined progressively into the earliest Ashgill before increasing again toward the end of Ordovician time (cf. Tuckey 1990).

Completeness of the Fossil Record One of several different methods (Paul 1998) of evaluating the completeness of the fossil record is gap analysis in which the proportion of Lazarus taxa per time interval is calculated. Figure 16.7 depicts the results for stenolaemate genera through the time slices of the Ordovician. An increasingly high proportion of Lazarus taxa characterizes the later part of the Ordovician, indicating a poor fossil record, culminating in TS.6c, where more than half of the genera known to be present by range extrapolation are missing from the record. This finding implies that the decline in bryozoan diversity during the late Caradoc to Ashgill (figures 16.1 and 16.3) is partly due to deficiencies in the fossil record. A mediocre record is also apparent

0.6 Proportion of Lazarus Genera 0.5

0.4

0.3

0.2

0.1

0 3a

3b

4a

4b

4c

5a

5b

5c

5d

6a

6b

Time Slice

FIGURE 16.7. Proportion of Lazarus genera of bryozoans for TS.3a to 6c (see figures 16.1, 16.3, 16.4).

6c

for TS.3a–4a of the early to mid Arenig, although the lower diversity here means that less significance can be attached to this finding. ■

Discussion Patterns of Radiation

Three earlier papers have depicted the pattern of bryozoan generic radiation during the Ordovician (Taylor and Larwood 1990: figure 10.2; Anstey and Pachut 1995: figure 8.5; Sepkoski 1995: figure 2). Generic range charts are also given by Ross and Ross (1996), but diversities are not tallied. These analyses, employing overlapping global databases and varying in their stratigraphic resolution, reveal patterns almost identical to the pattern shown here in figure 16.4. All show a prolonged diversification for most of the Ordovician before a diversity decline at the end of the period, although the Sepkoski (1995) curve shows an unexpected second peak in diversity at the very end of the Ordovician. Species-level assessment of Ordovician diversity changes has not previously been attempted at this stratigraphic precision (cf. Horowitz and Pachut 2000). Our results indicate that the increase in diversity from mid Arenig to mid Caradoc followed an almost perfectly exponential trajectory (figure 16.3), suggesting unconstrained diversification. If diversification was truly exponential, it can be predicted that stenolaemate bryozoans were present as early as the first or second Ordovician time slices (TS.1a or 1b) in the Tremadocian, which is where the line of regression of figure 16.3 intersects the y-axis. This offers hope for the future discovery of skeletally preserved bryozoans following a more thorough searching of rocks of this age. The leveling-off and decline in diversity at the end of the Ordovician can be viewed in the contexts of (1) a worsening fossil record as revealed by gap analysis (figure 16.7); (2) the Late Ordovician mass extinction (see Tuckey and Anstey 1992); and (3) the beginnings of a diversity plateau that characterized the remainder of the Paleozoic (Taylor and Larwood 1990: figure 10.1), arguably indicating attainment of the equilibrium diversity level identified by Sepkoski (1979) for marine families as a whole. Several authors have remarked on the emergence of most higher taxa (orders and suborders) of bry-

Bryozoans ozoans during the Ordovician (e.g., Taylor and Curry 1985; Anstey and Pachut 1995; Markov et al. 1998). Hierarchical taxonomic structure alone is capable of explaining why diversification at high taxonomic levels occurred so early in the evolution of the phylum. The sudden appearance of higher taxa is paralleled by numerous evolutionary innovations, including taxon-defining apomorphies. The subsequent evolution of Paleozoic stenolaemate bryozoans is to a large extent based on recombination (Markov et al. 1998) and/or parallel origin of characters first seen in the Ordovician, hence the high levels of convergence exhibited by Paleozoic bryozoans (e.g., Anstey and Pachut 1995). Most of the major colony growth forms found among Paleozoic bryozoans appeared during the Ordovician radiation of the phylum (Larwood and Taylor 1979), and their main ecologic roles were also established at this time (Ross 1984). There are clear similarities between the bryozoan radiation pattern and those published for other groups, for example, “articulate” brachiopods, echinoderms, corals, gastropods, bivalves, and conodonts (see Sepkoski 1995 and chapters in this volume). Although some of these groups (e.g., brachiopods, echinoderms, conodonts) commenced radiation before the Bryozoa, they all show a sustained pattern of diversity increase until the Caradoc or Ashgill, followed by a diversity drop close to or at the end of the Ordovician.

Processes Driving the Radiation According to Sepkoski and Sheehan (1983), there is no obvious physical trigger, such as major sea level change, for the Ordovician Radiation of marine animals. The advent of “Calcite Seas” in the Early Ordovician facilitated the rapid formation of carbonate hardgrounds and would have increased ecologic habitats and evolutionary opportunities for bryozoans, echinoderms, and other benthic organisms utilizing hard substrates for attachment (Wilson et al. 1992; Rozhnov 2001). Whether this was sufficient to trigger explosive evolutionary radiation in bryozoans is contestable, however. Ross (1985) suggested that climatic changes may have been involved in the Ordovician radiation of bryozoans, with diversity increasing when the climate became warmer and declining as it became cooler, particularly during the Ashgill. Anstey

(1986) hypothesized that the loss, related to global cooling and eustatic lowering of sea level, of a highly diverse region—the Cincinnati Province—was the chief factor in the extinction of bryozoans during the Late Ordovician, at least within North America. Eustatic sea level changes have also been implicated more widely in Ordovician (and Paleozoic) diversity patterns of bryozoans by Ross and Ross (1996). Transgressions led to diversity increases, as in the early Caradoc, and regressions to decreases, as at the end of the Ashgill. The post-Tremadocian match between the diversity pattern found here (figures 16.1 and 16.4) and the sea level curve published by Ross and Ross (1996: figure 1) is not exact. Although the Caradoc-Ashgill patterns are well correlated, the low sea levels that pertained during most of the Arenig and throughout the Darriwilian correspond with a marked phase of bryozoan diversification. The control of bryozoan diversity by an extrinsic factor or factors during the Ordovician is attractive in view of the parallel patterns exhibited by other phyla, but it is difficult to ignore the role of intrinsic changes in driving the radiation of the Bryozoa. The acquisition of a mineralized skeleton was of profound importance to the Bryozoa in paving the way for morphological and ecologic expansion not possible in primitive, soft-bodied bryozoans. This key evolutionary innovation may have been sufficient to trigger the seemingly unconstrained exponential diversification of bryozoans that persisted until mid Caradoc times (figure 16.3). Analogies with cheilostome bryozoans, which radiated explosively in the Cretaceous, point to another possible key evolutionary innovation (Taylor and Larwood 1990). Cheilostome radiation commenced immediately after the first skeletal evidence for larval brooding, suggesting a macroevolutionary model whereby the advent of short-lived brooded larvae led to decreased gene flow between populations and a dramatic increase in the rate of allopatric speciation (Taylor 1988). Unfortunately, this model is not testable in stenolaemate bryozoans until skeletal criteria for distinguishing brooding species have been established.

Future Research Many questions remain to be addressed regarding the Ordovician radiation of bryozoans. These demand

155

156

   .                  further analysis of already published data as well as new empirical studies of fossil bryozoans. A fuller appreciation of macroevolutionary patterns can be obtained by looking at changes in assemblage diversity and biomass, as has been done by McKinney et al. (2001) for post-Paleozoic bryozoans. Research is needed to see whether these changes correlate with or are decoupled (cf. Droser et al. 2000) from the global radiation pattern found here. Further cladistic analyses are required of Ordovician bryozoans to define clades with more confidence and precision. It will then be possible, for example, to study the dynamics of individual clades during the Ordovician radiation and to study temporal, biogeographic, and environmental patterns of parallel character acquisition in different clades. Quantification of disparity changes among bryozoans during the Ordovician has yet to be undertaken. Such research on morphospace occupancy would provide an additional perspective on the Ordovician radiation of bryozoans, as well as a potentially instructive comparison with published findings on the radiation of a noncolonial phylum (Arthropoda) during an earlier (Cambrian) major radiation (Fortey et al. 1996). Concerning possibilities of new fossil discoveries, extrapolation of the exponential species diversification curve (figure 16.3) predicts that stenolaemate bryozoans should occur back to the base of the Ordovician. Carbonate platforms of Tremadocian age, not unlike those known to be inhabited by bryozoans later in the Ordovician, have so far failed to produce any definite records of bryozoans. It may be that the earliest stenolaemates lived in more nearshore clastic environments and migrated offshore with time, in accordance with a general pattern described for the Ordovician by Sepkoski and Sheehan (1983). Clastic facies are rarely studied for bryozoans because of the typically decalcified preservation of their shelly faunas. Importantly, the oldest known occurrence of bryozoans outside China is a decalcified esthonio-

porine from nearshore clastics of early Arenig age in Wales (Taylor and Cope 1987). The most primitive stenolaemates, corynotrypid cyclostomes, are less likely to be noticed because of the small size of their branched, encrusting colonies. ■

Conclusions

The Bryozoa are a diverse phylum of benthic, colonial, suspension-feeding metazoans. They have no unequivocal Cambrian fossil record—the oldest fossil bryozoans are recorded from the Tremadocian. A major radiation of bryozoans occurred during the Ordovician, with the rapid appearance of all but one of the orders of marine bryozoans generally recognized in the Phanerozoic, the attainment of a family diversity level that was maintained as a plateau for the remainder of the Paleozoic, and a fossil record comprising more than 1,000 described species. Acquisition of a calcareous skeleton was very important, not merely for providing fossilizable hard parts but in allowing morphological and ecologic expansion of the phylum. Analysis of taxic patterns of diversity change and turnover through the Ordovician based on 19 time slices (each between 1.5 and 3 m.y. in duration) shows an exponential increase in species diversity from mid Arenig to mid Caradoc, followed by a species (and genus) diversity decline that is particularly pronounced at the very end of the Ordovician, where deficiencies in the fossil record are implied by the high proportion of Lazarus genera. Some variations in the pattern of radiation are evident according to major taxonomic group and biogeographic province.

 We thank Barry Webby for inviting us to participate in IGCP 410 and the Deutsche Forschungsgemeinschaft (DFG) for funding one of us (AE) through a fellowship.

17

Brachiopods David A. T. Harper, L. Robin M. Cocks, Leonid E. Popov, Peter M. Sheehan, Michael G. Bassett, Paul Copper, Lars E. Holmer, Jisuo Jin, and Rong Jia-yu

T

he Ordovician brachiopod radiation was the most marked interval of diversification in the entire history of the phylum (Harper and Rong 2001). Apart from the initial acquisition of hard shells in the earliest Cambrian, the Ordovician was the most important period for brachiopod evolution and diversification in the whole Phanerozoic with the continued exponential increase in numbers of brachiopod genera seeded during the Cambrian (Patzkowsky 1995a). A series of stepwise radiations across most of the major orders helped set the agenda for much of life on the Paleozoic seafloor. In particular, the articulated brachiopods, now termed the subphylum Rhynchonelliformea, diversified from the 4 superfamilies Billingselloidea, Orthoidea, Plectorthoidea, and Porambonitoidea present at the end of the Cambrian into the 19 superfamilies (Eichwaldioidea, Strophomenoidea, Plectambonitoidea, Chonetoidea, Chilidiopsoidea, Triplesioidea, Skenidioidea, Orthoidea, Plectorthoidea, Dalmanelloidea, Enteletoidea, Porambonitoidea, Pentameroidea, Stricklandioidea, Ancistrorhynchoidea, Camarotoechioidea, Atrypoidea, Athyridoidea, and Cyrtinoidea) all present by the end of the period; only the Billingselloidea, Polytoechioidea (Tremadocian to Caradoc), and Clitambonitoidea (Arenig-Ashgill) became extinct within the Ordovician. By the end of the Ordovician, with the exception of the spinose productoids, superrecumbent strophalo-

sioids, the coral-like richthofenioids, and the bizarre lyttonioids, the majority of morphological innovations had already appeared (Harper and Wright 1996). Key morphological adaptations promoted both taxonomic variety and a spectrum of lifestyles despite the relatively constrained skeletal architecture of the brachiopod animal. The radiation progressed with the expansion of brachiopod clades into outershelf and slope settings together with carbonate platform and reef environments. Such colonizations involved particular specializations (Copper 2001a). Nevertheless, across the phylum macroevolutionary changes were apparently decoupled from fluctuations in taxon abundance and taxonomic diversity (Droser et al. 1997); moreover, ecologic changes were not necessarily linked to crashes or hikes in taxonomic diversity (Droser et al. 2000). The origins for the radiation are unclear, in part owing to the time lag between the appearance of macroevolutionary innovations and subsequent apparent diversifications. The role of marginal and intraoceanic sites, many located within Early Paleozoic mountain belts or destroyed by subduction-related processes, may have been important during the early stages of the radiation (Harper and Mac Niocaill 2002). During the radiation, a shift in ecologic bias occurred during the early Darriwilian, particularly on cratons such as Laurentia, from communities with abundant trilobites to those dominated by brachiopods 157

158

     . .          . (Droser and Sheehan 1997b). Bambachian megaguilds were filled, and new paleocommunities were generated as the radiation progressed (Bottjer et al. 2001). ■

Taxonomic Diversification

The taxonomic component of the diversification involved an expansion of taxa at all levels, but the escalation is most marked at the generic level. For example, nearly 30 articulated genera are known from the lower Tremadocian, whereas a total of 185 genera have been recorded from lower to middle Ashgill horizons. Nevertheless, the radiation was not uniform, with peaks and troughs varying considerably across the different superfamilies. The maximum diversity of individual assemblages increased significantly from fewer than 10 taxa in the Upper Cambrian to more than 30 genera in many Late Ordovician paleocommunities. The diversification of the nonarticulates is less obvious; nonetheless, more than 150 Ordovician genera are known.

Phylogeny A relatively stable, cladistically based classification (Williams et al. 1996) has been developed within the framework of the revised Treatise volumes for the phylum (Kaesler 2000). The broad-frame phylogeny recognizes three subclasses: the linguliformeans, the craniiformeans, and the rhynchonelliformeans. Generally the first two divisions include the “inarticulates” of previous classifications, whereas the last division embodies the concept of the “articulate” brachiopod. The Ordovician Radiation involved “nonarticulated” Lingulida, Acrotretida, and Siphonotretida together with the Craniopsida, Craniida, and Trimerellida; the paterinates were present during the event but did not contribute significantly to the Ordovician biodiversifications. The “articulated” groups included the Strophomenida, Triplesiidina, Billingsellidina, Clitambonitidina, Protorthida, Orthida, Pentamerida, Rhynchonellida, Atrypida, and Athyridida; the Productida (Chonetoidea), Orthotetida, and Spiriferida were present at low diversities prior to radiations during the Silurian, whereas the enigmatic Dictyonellida were a very minor part of the later Ordovician brachiopod fauna.

Approach and Methodology The database has been developed from a series of range charts. For virtually all the groups discussed, up-to-date information is available through research associated with the revised volumes of part H (Brachiopoda) of the Treatise on Invertebrate Paleontology. In this respect both brachiopod phylogeny and taxonomy are state of the art. A stable stratigraphic framework is provided by the continuing researches of the IUGS Subcommission on Ordovician Stratigraphy. Nevertheless, there are, of course, error limits in both range and geographic data. The known stratigraphic distributions of all genera have been plotted within a framework of 19 chronostratigraphic and geochronologic intervals developed for the Ordovician System (e.g., figure 17.1). Standing diversity has been plotted for most of the major groups together with first and last appearance information. Moreover, for some groups diversity counts for each stratigraphic interval have been corrected. The main trends, however, are clearer when the raw data are displayed graphically. The “endpoint correction” inserts the value of 0.5 for a “first appearance datum” (FAD) or a “last appearance datum” (LAD) in a given sample, whereas a value of 0.33 is inserted for the occurrence of both a FAD and a LAD. Corrections were calculated with PAST software (Hammer et al. 2001) (http://folk.uio.no/ohammer/past/). The genus origination/extinction rates, plotted as per lineage million years (Lma), were calculated as the number of generic originations/extinctions within a given stratigraphic interval, divided by the total generic diversity within the interval, divided by the chronologic duration of the corresponding time interval (see Patzkowsky and Holland 1997 for methodology and discussion).

Linguliformean and Craniiformean Brachiopods (LEP, LEH, MGB) The linguliformean and craniiformean brachiopods together represent a relatively numerically insignificant component of the Ordovician fauna as a whole. Nevertheless, these two groups formerly regarded as “inarticulates” are each rather different in their environmental requirements and dispersal potential, defined mainly by differences in their larval

Brachiopods

Number of Genera

FIGURE 17.1. Ordovician standing diversity patterns of the linguliformeans, craniiformeans, and rhynchonelliformeans in terms of numbers of genera.

1c–1d

2b–2c

development. Linguliformean genera with their planktotrophic larval phases usually show a more widespread and even biogeographic distribution during the Ordovician mostly controlled by climatic factors and similar to the distributional patterns of pelagic animals such as conodonts, whereas craniiformeans with their lecithotrophic larvae that fed on yolk were probably unable to cross even narrow oceanic tracts, and their patterns of biogeographic distribution and diversification in Ordovician depended strongly on the relative position of the ancient continents. The linguliformeans represent, however, together with trilobites, the two most distinctive benthic groups of the Cambrian Evolutionary Fauna (EF) (Sepkoski 1981a). The Cambrian record of the craniiformeans is very sparse (Popov et al. 1999b). Only two craniopside genera were reported from the Early and Mid Cambrian, followed by a gap in known stratigraphic distributions until their reappearance in the Late Cambrian and Tremadocian. All three orders of craniiformeans (Craniida, Craniopsida, and Trimerellida) are characterized by the lack of a pedicle at all growth stages and instead pursued a free-lying, cemented, or encrusting life habit. Larvae of the Recent genus

6a–6b

Novocrania have a free-swimming stage of very short duration, and this condition had probably evolved very early in the history of group, as is evident from the biogeographic patterns and life strategies of the group already established in the Ordovician. Major biodiversification and biogeographic expansion of the group have occurred since the Mid Ordovician (Popov et al. 1999b); thus they, in contrast to the linguliformeans, have a diversity history similar to that of the Paleozoic Evolutionary Fauna (EF). Linguliformea The Lingulida and Acrotretida had already achieved considerable evolutionary success during the initial Cambrian radiation of the metazoans and were common in all types of marine environments by the end of the Cambrian. Their Ordovician record includes 150 genera (Lingulida—85 genera; Acrotretida—47 genera; Siphonotretida—15 genera; and Paterinida— 3 genera). Geologic time-frequency distributions showing plots of generic diversity within linguliformean orders, together with the numbers of first and last appearances of Ordovician genera, are depicted in figure 17.2. Origination and extinction rates

159

     . .          . Lma 0.40

Subphylum Linguliformea

Lma Order Lingulida

0.30

extinction rate origination rate

0.30

extinction rate origination rate

0.20

0.20 0.10

0.10

4c

5a

5b

5cÐd

6aÐb 6c

Upper

Ordovician

IaÐb

2a

IcÐd

3aÐb 4aÐb

4c

5a

450 5b

5cÐd

6aÐc 6c

Upper

Middle

Lower

440 Silurian

4a-b

Middle

460

470

480 2bÐc

3aÐb

Age Ma 489

440

Upper Cambrian

2a

IcÐd Lower

450

Silurian

IaÐb

460

470

480 2bÐc

Upper Cambrian

Age Ma 489

Ordovician

Lma Order Acrotretida

Lma

Order Siphonotretida

extinction rate

0.30

extinction rate

origination rate

origination rate

0.20

0.20

0.10

3aÐb 4aÐb

4c

5a

Middle

5b

5cÐd

Upper

6aÐb 6c

Ordovician

Age 489 Ma

IaÐb

IcÐd Lower

460

470

480

2a

3aÐb

4aÐb

4c 5a

Middle

450 5b

5cÐd

Upper

440 6aÐb6c

Silurian

2a

440

2bÐc

IcÐd Lower

450

0.10

Upper Cambrian

IaÐb

460

470

480

Silurian

489

2bÐc

Age Ma

Upper Cambrian

160

Ordovician

FIGURE 17.2. Extinction and origination rates (clockwise from top left) of the linguliformeans, lingulides, siphonotretides, and acrotretides (after Bassett et al. 1999b); see text for calculation of the rates (Lma).

for genera of the Lingulida, Siphonotretida, and Acrotretida are also plotted in figure 17.2. There is a considerable decline in generic diversity in the second half of the Late Cambrian from the Proconodontus Zone (Bassett et al. 1999b; Holmer et al. 2001). Linguliformean faunas on the outer shelf were most affected, and more than 90 percent of the Late Cambrian genera became extinct by the beginning of the Ordovician. In the early Tremadocian (TS.1a–b; see time slices in chapter 2), acrotretide diversity of the microbrachiopod assemblages was reduced to one to three genera, and the Broggeria fauna, adapted to dysaerobic environments, spread across high and temperate latitudes following an expansion of the areas of black shale deposition (Popov and Holmer 1994, 1995). Recovery of linguliformean faunas on all the major Early Paleozoic platforms occurred during the late Tremadocian–early Arenig (TS.1b–2c, or equivalent to the Paltodus deltifer to Oepikodus evae conodont zones), and by the mid Arenig (TS.3a) their generic

diversity had been restored to the maximum observed in the Cambrian. The recovery characteristics were different for the faunas that inhabited the nearshore and outer-shelf environments. The distinctive feature of the newly emerging microbrachiopod assemblages, which spread widely across the outer shelf and basin, during the later Ordovician is abundance of the acrotretide families Ephippelasmatidae, Eoconulidae, Scaphelasmatidae, Torynelasmatidae, and Biernatidae together with the lingulides Rowellella, Elliptoglossa, and Paterula and siphonotretides (Popov and Holmer 1994, 1995). Environments inhabited by these linguliformean faunas also developed as refugia for the last paterinides, where they survived until the terminal Ordovician extinction. Presently available data from Kazakhstanian and the Central Asian terranes (Holmer et al. 2000; Popov and Holmer 1995) suggest that the Ephippelasmatidae, Eoconulidae, Scaphelasmatidae, and possibly Torynelasmatidae, which constitute the core of the Ordovician micromorphic brachiopod faunas,

Brachiopods evolved first in an island-arc setting located in the temperate and low latitudes between Baltica and East Gondwana. The first pulse of diversification of the micromorphic linguliformeans in the outer shelf is recorded in the late Tremadocian (TS.1c), when this fauna suddenly appeared on the opposite sides of Baltica (Baltoscandian basin and South Urals). They sustain a second pulse of generic biodiversification during the mid Arenig (TS.2c), when they are recorded for the first time in Laurentia (Krause and Rowell 1975) and South China (Zhang 1995); but in general their taxonomic structure at the family level and generic diversity remained relatively stable until the mid Ashgill (TS.6a–b) (Wright and McClean 1991; Bassett et al. 1999b; Popov 2000b). The diversity of the lingulide family Elkaniidae, which was most characteristic of outer-shelf environments during the Early and Mid Ordovician, had decreased by the Late Ordovician to the single genus Tilasia, which survived until the mid Ashgill. The terminal Ordovician extinction strongly affected the composition of the micromorphic brachiopod assemblages at the generic level, but it is not so clear at the family level. Among acrotretide families only Eoconulidae and Ephippelasmatidae became extinct, but a number of Silurian genera belonging to other families (Acrotretidae, Biernatidae, Scaphelasmatidae, and Torynelasmatidae) were reduced dramatically, and each family is represented by only a single genus in the Silurian. Although the siphonotretides survived the terminal Ordovician extinction, they became very rare in Silurian and Devonian rocks (Mergl 2001). During the Mid and Late Cambrian, linguliformeans, especially the obolids and zhanatellids, became widespread across the shallow clastic shelves. In particular, they developed adaptations that permitted colonization of mobile sandy substrates, where they retained their dominant position through the Early Paleozoic. These nearshore lingulide faunas were partly affected by regional extinctions in Baltica and Laurentia near the Cambrian-Ordovician boundary. It is also likely that the low- to temperate-latitude lingulide faunas of West Gondwana (including North Africa, the Middle East, Avalonia, and the Perunica microcontinent [= Bohemia]) played an important role in the origin of the Ordovician shallow-shelf assemblages. A good example is the Thysanotos-

Leptembolon fauna, which probably evolved in some peri-Gondwanan settings and soon spread across Baltica (Popov and Holmer 1994; Mergl 1997). High origination rates of the lingulides and siphonotretides during the late Tremadocian (TS.1d) were caused mainly by the proliferation of the shallow-shelf assemblages (figure 17.2). The Darriwilian (TS.4a–c) was the interval of significant reorganization of shallow-shelf assemblages with increased turnover rates (figure 17.2), when epibenthic lingulides were completely replaced by taxa with a burrowing habit. It occurs well before the first appearance of bivalves on the clastic shelves of the Kazakhstanian terranes and the shallow carbonate shelves of Baltica; thus, it is rather unlikely that increased bioturbation promoted by the colonization of bivalves is responsible for the extinction of the epibenthic lingulide and siphonotretide taxa. During that time some shallow-shelf lingulide assemblages (e.g., the Hyperobolus-Talasotreta association) indicated a considerable shift toward basinal environments (Holmer et al. 1996), but in general this change is not characteristic of the Ordovician linguliformean faunas. Despite a persistence of the family stocks, somewhere between 50 and 60 percent of the total number of genera disappeared from linguliformean assemblages during the Llanvirn and early Caradoc (TS.4a–c, 5a) (figures 17.2, 17.3). The number of siphonotretides dropped to six genera at the beginning of the Llanvirn (TS.4a–b) (figure 17.2). In nearshore and shallow-shelf environments, faunas of epibenthic lingulides of the families Obolidae and Zhanatellidae were replaced completely by lingulidemollusk assemblages, in which lingulides, adapted to life in burrows, became the predominant brachiopods (Bassett et al. 1999b). Most of the evolutionary novelties that defined morphological and anatomic characters of the orders and superfamilies, as well as major adaptations, evolved in linguliformeans well before the beginning of the Ordovician. The only exception is the first appearance of the superfamily Discinoidea in the early Arenig, which represents the last significant evolutionary event in lingulide history. In contrast with the majority of other lingulides, they acquired a combination of features such as an elongate pedicle track posterior to the ventral umbo covered anteriorly by a listrium, holoperipheral growth of the ventral valve (family Trematidae)

161

162

     . .          . or both valves (family Discinidae), and complete reduction of pseudointerareas, apart from a narrow strip retained occasionally in some early genera. In the acrotretides the most significant morphological novelty was the elaboration of the dorsal median septum, involving the development of spines and widely variable morphologies of the surmounting plate that evolved in several genera of the acrotretide families Acrotretidae (e.g., Fascicoma), Ephippelasmatidae (e.g., Numericoma), and Torynelasmatidae (e.g., Cristicoma and Polylasma) during the mid to late Arenig and early Llanvirn (TS.3c–4b) (Holmer and Popov 2000; Nazarov and Popov 1980). Adaptation to a burrowing life mode possibly evolved sometime in the Late Cambrian from an escape strategy within the epibenthic lingulides that inhabited mobile sandy substrates (Popov et al. 1989). This shell morphology, as well as the burrowing habit, probably evolved convergently and slightly diachronously in a number of lingulide lineages, since taxa with both finely pitted and smooth larval and postlarval shells are known. However, most of the characteristic features of the infaunal lingulides—for example, convergent vascula lateralia in both valves and the complete reduction of the dorsal vascula media—evolved later, sometime in the Llanvirn. During the Ordovician radiation of linguliformeans, most of the first-order morphological innovations occurred during the Early Ordovician. From the Arenig onward, linguliformeans were insignificant components of benthic assemblages, which became dominated by filtrators such as the rhynchonelliformean brachiopods, bryozoans, and echinoderms. However, some linguliformeans continued to form distinctive parts of the benthos in marginal environments such as eutrophic basins, mobile sands on shallow shelves, and abyssal basins. But in general, post-Arenig linguliformeans had much lower, nearbackground genus origination rates (0.04–0.07 Lma) compared with the other major groups of Sepkoski’s (1981a) Paleozoic EF, including, for example, the rhynchonelliformean brachiopods (Patzkowsky and Holland 1997). Craniiformea The oldest known Ordovician craniiformeans, not yet formally described, are those from the lower

Arenig of South Tien Shan in Central Asia, where they are represented already by two craniid genera including an encrusting Petrocrania associated with the rhynchonelliformean brachiopods Clarkella and Tritoechia and a linguliformean assemblage very similar to the early Arenig assemblages (TS.2a–b) of Baltica (Holmer et al. 2000). Another possible island-arc setting where craniides already occur during the Arenig is the Hinggan Ling Mountains of northeastern China, where Celidocrania occurs together with other late Arenig brachiopods of mixed Avalonian and Baltoscandian affinity (Liu et al. 1985; Neuman and Harper 1992). Craniides had already colonized Baltica in the mid Arenig (TS.3a–b or Volkhov regional stage), where they are represented by Pseudocrania and Petrocrania. There is also a report on the occurrence of a craniopside from the lower Llanvirn of Öland, Sweden. The environments inhabited by craniides at that time were located within shallow pericontinental seas (BA 3 and 4), associated with storm-generated sedimentation and widespread hardgrounds. During the “Llandeilo” (TS.4c–5a, Husteograptus teretiusculus and Nemagraptus gracilis graptolite zones), craniides diversified in Baltica to four genera. About the same time the earliest-proven craniopside Pseudopholidops became common in the Kukruse stage, and also the earliest trimerellides emerged in an islandarc setting between Baltica and East Gondwana, now incorporated in the Kazakhstanian orogen. The occurrence of three trimerellide genera in ChingizTarbagatay is of particular significance. Here, for the first time, they formed a prominent nucleus of lowdiversity/high-density shallow marine (BA 3) benthic assemblages inhabiting carbonate sands and muds in close proximity to carbonate mud mounds. At present there is no pre-Caradoc Ordovician record of craniiformeans from mainland Gondwana, Laurentia, Avalonia, and Siberia. The early Caradoc (TS.5a) was the time of craniiformean diversification; taxa spread widely across the temperate and low latitudes (figure 17.3). There was a marked radiation of trimerellides in East Gondwana (Australia), and Eodinobolus reached Laurentia. In both regions they became dominant taxa in low-diversity, shallow-water (BA 2–3) assemblages (Norford and Steele 1969; Percival 1995). Such faunas remained notably absent in high-latitude West Gondwana.

Brachiopods

Subphylum Linguliformea

Number of Genera 60

Paterinida & Siphonotretida

50 40

Lingulida

Acrotretida

30 20

first appearances

10

last occurrences

480

Lower

460

450

440

3aÐb 4aÐb 4c 5a 5b 5cÐd 6aÐb 6c

Middle

Ordovician

Upper

Silurian

IaÐb IcÐd 2a

470 2bÐc

Upper Cambrian

Age 489 Ma

Number of Genera

Subphylum Craniiformea

20 15

Craniopsida Trimerellida

10

Craniida

IaÐb IcÐd 2a

Lower

470

460

450

440

3aÐb 4aÐb 4c 5a 5b 5cÐd 6aÐb 6c

Middle Ordovician

Upper

Silurian

480 2bÐc

Upper Cambrian

Age 489 Ma

5

Lma

Subphylum Craniiformea origination rate

0.3

extinction rate

0.2 0.1

IaÐb IcÐd 2a

Lower

470

460

450

440

3aÐb 4aÐb 4c 5a 5b 5cÐd 6aÐb 6c

Middle Ordovician

Upper

Silurian

480 2bÐc

489

Upper Cambrian

Age Ma

FIGURE 17.3. From top, absolute abundance of linguliformean orders (after Bassett et al. 1999b), absolute abundance of craniiformean orders, and extinction and origination rates across the craniiformeans (after Popov et al. 1999b).

From the early to mid Caradoc (TS.5b), craniides and craniopsides appeared in Laurentia, where they became relatively common in shallow marine biofacies (BA 2–3) on carbonate platforms. By that time

craniides also reached Bohemia (Petrocrania), and craniopsides (Paracraniops) and craniides (Orthisocrania) with strong Baltic affinities appeared in Avalonia. The late Caradoc–early Ashgill (figure 17.3; TS.5c–d, 6a–b) was an interval of maximum Ordovician radiation and expansion of the craniiformeans. Trimerellides spread from Kazakhstanian terranes toward Baltica and Siberia (Popov et al. 1997). Diverse trimerellide faunas appeared suddenly on the Yangtze Platform in the early Ashgill. Their ecology and origin are poorly known, but high endemicity at the generic level may suggest that they evolved over a considerable period of time in relative isolation (Popov et al. 1997). High-latitude Gondwana again remained unaffected by this late Caradoc–early Ashgill event. A significant decline in the diversity of craniiformean faunas took place in the mid Ashgill (upper TS.6a–b; graptolite-based Dicellograptus anceps Zone), affecting faunas across Kazakhstanian terranes, the Yangtze Platform, and East Gondwana. They disappeared completely from Siberia. Only the enigmatic Gasconsia suggests a pattern of geographic expansion in Kazakhstan and Baltoscandia, following the general distribution of the Holorhynchus brachiopod fauna. Extinction of Ordovician trimerellide faunas across Baltica, Laurentia, and the Kazakhstanian terranes took place during the Hirnantian. Among the craniopsides and craniides only Petrocrania and Pseudopholidops are relatively widespread as rare components of the Hirnantia brachiopod assemblages. Among the craniiformeans only the Trimerellida achieved considerable evolutionary success during the Ordovician, whereas the Craniopsida and Craniida remained insignificant parts of benthic assemblages dominated by filtrators. Separation of the Craniida and Trimerellida from the craniiformean stem group, which possibly consisted of craniopsides, was linked initially to adaptations to high-energy, shallowwater environments. The origin of an encrusting habit in the craniides made it possible for them to inhabit hardgrounds and rocky bottoms in such environments of the shallow shelf, which were otherwise hardly accessible to most Early Paleozoic benthic animals. The great increase in shell size in trimerellides also gave them the ability to colonize mainly shallow-water habitats affected by seasonal storms (Webby and Percival 1983). The ability to undertake

163

164

     . .          . migration was severely limited in Ordovician craniiformeans, and their appearance across various plates and on ancient island arcs was markedly diachronous (Popov et al. 1999b: figure 1). Craniopsides and craniides remain unknown from Siberia if we exclude the Taimyr Peninsula (Cocks and Fortey 1998). Their appearance in the Caradoc of Avalonia was close to the time of its collision with Baltica. The Ordovician occurrences of trimerellides suggest a distinct pattern of island-arc settlement and biogeographic isolation, as well as indicating the possible position of Laurentia in proximity to East Gondwana sometime during the Darriwilian to early Caradoc (Popov et al. 1997).

Rhynchonelliformea Strophomenata Strophomenida (LRMC). The two superfamilies within the order Strophomenida, the Strophomenoidea and the Plectambonitoidea, both have their first representatives in Ordovician rocks (Cocks and Rong 2000). The plectambonitoids were the earlier, perhaps originating from the Billingselloidea during the Tremadocian and differing from the latter principally in having a fibrous pseudopunctate shell. The Strophomenoidea have a pseudopunctate shell, but laminar with a bilobed cardinal process, and were probably derived from the Plectambonitoidea in the Arenig, with the earliest known strophomenoid from the South China plate. So far as is now known, 95 plectambonitoid and 80 strophomenoid genera and subgenera were present in the Ordovician (figures 17.4, 17.5). Both superfamilies survived the end Ordovician turnovers, although with initially reduced diversity, with the Plectambonitoidea becoming extinct in the Late Devonian at the Frasnian-Famennian boundary and the Strophomenoidea continuing on into the Late Carboniferous. Plectambonitoids appear to have originated in midshelf (BA 3) communities on the low-latitude terrane of the Altai-Sayan, with Akelina as the first known genus, and migrated to more temperate paleocontinents such as Baltica by the Arenig. In the latter half to two-thirds of the Ordovician, plectambonitoid genera expanded rapidly (figures 17.4, 17.5), reaching their all-time highest number of first appearances (FADs) in the early Llanvirn (TS.4b), and

in particular became well established and often even abundant in shallower-water (BA 2) communities as well as those farther offshore. They also spread across the paleolatitudes, with Aegiromena forming a notable constituent of the West Gondwanan and periGondwanan communities at high paleolatitudes. Strophomenoids radiated more slowly than plectambonitoids, reaching their first peak in the early Caradoc (TS.5a), and mostly occurred in midshelf (BA 3 and 4) communities throughout their history. A second peak of FADs in both superfamilies (and the highest for Ordovician strophomenoids) occurred in the early Ashgill (TS.6a). Last appearances (LADs) (figure 17.4) for the plectambonitoids had an early sharp peak in the early Llanvirn (TS.4b), probably reflecting the relatively large number of endemic genera known from that period, but the highest number of LADs was in the early Ashgill (TS.6a). The strophomenoids had a more even pattern of LADs, but with some peaking in the early part of the late Caradoc (TS.5c) and during the early Ashgill (TS.6a). Only four strophomenoid genera are actually recorded from the latest Ordovician Hirnantian stage (TS.6c), but it is instructive to note that five Ordovician strophomenoid taxa are known as Lazarus taxa from the Silurian, even though none has yet been recovered from rocks of Hirnantian age. Similarly, although only seven plectambonitoids are known from the Hirnantian, further four Ordovician Lazarus taxa are recorded from the Silurian. It is noteworthy that, in the uppermost Ordovician rocks and continuing on into the Silurian, most plectambonitoids, for example, the abundant Eoplectodonta, thrived best in deeper-shelf (BA 4 and 5) communities and were comparatively rare onshore. Although relatively uncommon during the first half of the Ordovician (figures 17.4, 17.5), both the Plectambonitoidea and the Strophomenoidea radiated quickly during the second half and often dominated some of the assemblages and communities in which they are found. A maximum of 43 genera or subgenera are recorded from a single time slice for each of the superfamilies, during the mid Caradoc (TS.5b) for the Plectambonitoidea and the early Ashgill (TS.6a) for the Strophomenoidea. In the Silurian, plectambonitoids dwindled in both numbers and diversity, in contrast to the strophomenoids, whose denticulated forms underwent further substantial radiations.

Brachiopods

Number of Genera

Number of Genera

Plectambonitoids First Plectambonitoids Last

1c–1d

2b–2c

6a–6b 1c–1d

2b–2c

6a–6b

Number of Genera

Strophomenoids First Strophomenoids Last

1c–1d

2b–2c

6a–6b

1c–1d

2b–2c

6a–6b

FIGURE 17.4. Strophomenide diversity. A, absolute abundances of the two superfamilies of strophomenide brachiopod. B, extinction and origination rates across the strophomenides. C, first and last appearances of the plectambonitoideans. D, first and last appearances of the strophomenoideans.

Chonetida and Orthotetida (LRMC). Although themselves relatively minor components of the Ordovician faunas, the era witnessed the origins of two orders of brachiopods that became numerous and diverse in later rocks. The Chonetida is represented by the single genus Archaeochonetes from the Ordovician and has been found in the Ordovician only from

the middle Ashgill of Anticosti Island, Canada, although it is also recorded from the Lower Silurian. The origins of the Chonetida are obscure in detail, although they almost certainly lie within the order Strophomenida: the ancestor was probably a plectambonitoidean within the family Sowerbyellidae, a family that has a cardinal process and internal brachial

165

     . .          .

A

B 80

60

70

50

60 40

50

Diversity

Diversity

166

40 30

30 20

20 10 10 0

0 1a 1b 1c–1d

2a 2b–2c 3a 3b 4a 4b 4c 5a 5b

Tremadoc Ibexian

Arenig

TREMADOCIAN

LOWER ORDOVICIAN 489 Ma

Llanvirn Whiterockian

5c 5d 6a–6b 6c

Caradoc Mohawkian

M. ORDOVICIAN 460.5

2a 2b–2c 3a 3b 4a 4b 4c 5a 5b

Tremadoc Ibexian

Arenig

TREMADOCIAN

DARRIWILIAN

472

1a 1b 1c–1d

Ashgill Cincinnatian

LOWER ORDOVICIAN

UPPER ORDOVICIAN 443

489 Ma

Llanvirn Whiterockian

5c 5d 6a–6b 6c

Caradoc Mohawkian

Ashgill Cincinnatian

DARRIWILIAN

M. ORDOVICIAN 472

460.5

UPPER ORDOVICIAN 443

FIGURE 17.5. Comparison of standing diversity (A) and corrected diversity (B) of strophomenide brachiopods measured in numbers of genera.

valve septa and other structures comparable to those of early chonetides but that lacked the characteristic chonetide spines on the hinge line. Ordovician Orthotetida all possess impunctate shells (in contrast to Silurian and later genera within the order), and it is uncertain whether they evolved from impunctate orthoids or pseudopunctate strophomenoids. There are four genera within the orthotetid Chilidiopsoidea in the Ordovician, and these are first represented by Gacella, first known from the middle Caradoc Chu-Ili terrane in Kazakhstan and from the upper Caradoc of Laurentia and elsewhere. In addition, the enigmatic Eocramatiidae, which is here also included within the Orthotetida (in contrast to other workers, who have included it within the Plectambonitoidea), includes only two rather rare genera, Eocramatia, with first and last appearances in TS.4b, and the closely related Neocramatia, recorded only from TS.5d and 6a (Harper 1989). Thus the order did not diversify much in generic terms during the Ordovician, but it became much increased in abundance, with Coolinia being a common constituent of the late Ashgill Hirnantia Fauna, which had a very wide distribution. Like the strophomenoids, the orthotetids were chiefly to be found in the middle shelf, principally in BA 3 and BA 4. Coolinia and Fardenia both survived into the Silurian, during which the

order underwent considerable radiation (Williams and Brunton 2000). Triplesiidina (DATH). This small but distinctive group, characterized by an unusual cardinalia, has 13 Ordovician representatives (Wright 1993, 2000). On the basis of shell morphology and especially the possession of a laminar secondary shell fabric, the triplesiidines have been closely related to the orthotetidines within the strophomenide clade. The group shows an extraordinary morphological variability in shell outline, profiles, and external ornament. The earliest taxon, Onychoplecia, occurs in the lower Llanvirn (TS.4b) rocks of western Newfoundland within the Toquima–Table Head Province. Nevertheless, both Oxoplecia and Triplesia have been described from the upper Llanvirn (TS.4c) rocks of Wales (Avalonia). The group underwent moderate radiations during the Caradoc and Ashgill, becoming widespread but never abundant. Billingsellida (DATH). The order Billingsellida is probably the least stable of the main taxonomic groupings. The order currently contains the Billingselloidea, Clitambonitoidea, and the Polytoechioidea (Williams and Harper 2000a). Opinion is divided on whether the billingsellides are monophyletic (Williams et al. 1996) or polyphyletic (Popov et al. 2001).

Brachiopods Within the billingsellidines, only Billingsella, Cymbithyris, Eosotrematorthis, Kozhuchinella, and Xenorthis have Lower Ordovician records, with only Eosotrematorthis reported from the Arenig (Williams and Harper 2000a). The group was widespread but with most occurrences at low latitudes. Two superfamilies, the Clitambonitoidea and the Polytoechioidea (Wright and Rubel 1996; Rubel and Wright 2000), are recognized within the clitambonitidines. The morphological features of the clitambonitoids suggest derivation from a protorthoid ancestor such as Arctohedra, rather than from within the billingsellide stem group, whereas most members of the polytoechioids may have evolved directly from the billingselloids and developed a quite different biogeographic distribution. The Clitambonitidina, apart from the polytoechioids, were a dominantly Baltoscandian group with all genera confined to the Ordovician. A massive diversification of the clitambonitoidines at the base of the Arenig (TS.2a) was mainly restricted to the Baltic Province and involved the clitambonitoids, preferring the cool-water carbonate facies of the Baltoscandian basin. By the end of the Caradoc (TS.5d) only four genera remained, with two, Kullervo and Vellamo, having relatively widespread distributions. The majority of the polytoechioids appeared during the ArenigLlanvirn interval and with the notable exceptions of Protambonites and Tritoechia were most characteristic of low latitudes. The last representatives, Admixtella and Peritritoechia, continued into the lower Caradoc in Central Asia and South China, respectively. Rhynchonellata Protorthida (DATH). The superfamily Skenidioidea represents this small but important group during the Ordovician (Williams and Harper 2000b). Genera are characterized by the apomorphies of a free spondylium and brachiophore plates that converge to form a plate underlying the notothyrial margin; both features were prominent in Early Cambrian taxa but also are typical of the post-Cambrian skenidioideans (Williams and Harper 2000b). The group includes five Ordovician taxa, first appearing during the Arenig in the circum-Iapetus region and China. The skenidiids were never common and conspicuous elements of the Ordovician radiations, but during the later

Ordovician the group anticipated the deeper-water habitats preferred by the family during the Silurian. Orthida (DATH). The pedunculate Orthida together with the recumbent Strophomenida dominated the brachiopod faunas of the Ordovician. Both groups underwent significant and sustained radiations during the period (figures 17.6, 17.7), most typically associated with siliciclastic environments. The Orthida probably represents the “modern articulated” brachiopod (Williams et al. 1996), and during the period the group was associated with a spectrum of habitats from nearshore to deep-sea environments together with carbonate mudmound and reef facies. The order Orthida (Williams and Harper 2000c) has been divided into two suborders, the Orthidina (impunctate orthides) and the Dalmanellidina (punctate orthides). The first has a significant Cambrian history. The second appeared during the Arenig; its origin within and relationship to specific orthidines remain uncertain. The greater orthide group contains more than 300 genera grouped into 46 families; orthide history encompassed the entire Paleozoic, although few taxa occur during the later parts of the era. Radiations occurred as a series of stepwise diversifications simulating ecologic displacements by successive superfamilies during the Early Paleozoic (Harper and Gallagher 2001); peaks in diversification are matched by expansions in morphological disparity (Harper and Gallagher 2001). The orthides, however, attained a maximum diversity during the Ordovician, accelerated by the Ordovician radiation event (figures 17.6, 17.7); nearly 90 orthide genera, dominated by the orthoids and dalmanelloids, characterize the Caradoc interval (Harper and Gallagher 2001). The Orthidina, comprising the Orthoidea and Plectorthoidea, appeared first, probably during the Early Cambrian (Williams and Harper 2000c). By the late Tremadocian, nearly 20 genera of orthidines, dominated by the orthoids, were present, many occupying shallow-water siliciclastic environments at high to temperate latitudes. The plectorthoids, however, were more common at low latitudes, commonly associated with carbonate environments. The first major radiation of the group occurred during the early Darriwilian (TS.4a) with an escalation in the numbers of orthoids and plectorthoids and the

167

     . .           .

Number of Genera

Number of Genera

1c–1d

2b–2c

1c–1d

2b–2c

Orthidines—First Orthidines—Last

6a–6b

Number of Genera

168

Dalmanellidines—First Dalmanellidines—Last

6a–6b

1c–1d

2b–2c

6a–6b

FIGURE 17.6. Orthide diversity. A, absolute abundances of the two suborders of orthide brachiopod. B, extinction and origination rates across the orthides. C, first and last appearances of the orthidines. D, first and last appearances of the dalmanellidines.

development of the dalmanellidine superfamilies, the dalamanelloids and enteletoids. Data suggest that this was the largest and most profound diversification with a standing count of 50 genera. Analysis of first and last appearances within the group nevertheless suggests that many were relatively short-lived stocks (figure 17.6C, D). Diversity again peaked during the mid Caradoc, with more than 50 genera. The dominant components

of this event were the dalmanellidines, particularly the dalmanelloids. Although a significant number of genera (16) appeared during the early Caradoc (TS.5a), the dominance of the group during the mid Caradoc is accentuated by a further diversification of both dalmanelloids and enteletoids and an increased decline among the impunctate groups. The dalmanellidines (Harper 2000) maintained their position during the later Ordovician, reaching

Brachiopods

B

A

50

70 60

40

Diversity

Diversity

50 40 30

30

20

20 10 10 0

0 1a 1b 1c–1d

2a 2b–2c 3a 3b 4a 4b 4c 5a 5b

Tremadoc Ibexian

Arenig

TREMADOCIAN

LOWER ORDOVICIAN 489 Ma

Llanvirn Whiterockian

5c 5d 6a–6b 6c

Caradoc Mohawkian

M. ORDOVICIAN 460.5

2a 2b–2c 3a 3b 4a 4b 4c 5a 5b

Tremadoc Ibexian

Arenig

TREMADOCIAN

DARRIWILIAN

472

1a 1b 1c–1d

Ashgill Cincinnatian

LOWER ORDOVICIAN

UPPER ORDOVICIAN 443

489 Ma

Llanvirn Whiterockian

5c 5d 6a–6b 6c

Caradoc Mohawkian

Ashgill Cincinnatian

DARRIWILIAN

M. ORDOVICIAN 472

460.5

UPPER ORDOVICIAN 443

FIGURE 17.7. Comparison of standing diversity (A) and corrected diversity (B) of orthide brachiopods measured in numbers of genera.

near parity with the orthidines during the early to mid Ashgill (TS.6a–b), when a number of short-lived taxa, signaled by increased first and last appearances, dominated the punctate orthides (figure 17.6C, D). During the Hirnantian (TS.6c) diversity dropped to fewer than 20 genera, matching diversities during the late Tremadocian. Assemblages were characterized by punctate orthides, the impunctate forms never recovering their dominant role after the end Ordovician extinction event. Pentamerida ( JJ). The Ordovician was marked by the presence of two quite different pentameride groups (Carlson 1996). During the Early and Mid Ordovician only the Syntrophiidina were present. The first Pentameridina did not evolve until the Ashgill. The largest radiation occurred in the order (figure 17.8) during the Tremadocian (TS.1a–c); only four genera, Bobinella, Mesonomia, Plectostrophia, and Palaeostrophia, continue through from the underlying Cambrian. There are more modest radiations during the early Arenig, Llanvirn, and mid Ashgill. Tremadocian faunas show high rates of origination and relatively low extinction rates. High rates of origination continue through the early Arenig, followed by low-diversity, holdover faunas during the mid Arenig. The most significant turnover, however,

occurs at the Arenig-Llanvirn boundary (TS.4a–b transition), when only four taxa continue over the boundary. The Caradoc (TS.5a–d) was characterized by relatively low diversities and a steady state of origins and extinctions. In contrast, the low-diversity Ashgill (TS.6a–c) pentameride fauna suffered high rates of extinction and was marked by a relatively large number of short-lived, endemic genera. Pentameride genera were generally cosmopolitan during the Tremadocian to the early Caradoc. In contrast, Ashgill taxa exhibit much greater degrees of provinciality. Rhynchonellida ( JJ). The rhynchonellide brachiopods first appeared during the Darriwilian (TS.4a–c) as low diversity faunas characterized by small shells occurring in the cool-water seas of the continental margins (Bassett et al. 1999c). Nevertheless, by the latest Darriwilian (TS.4c) the group had appeared in a variety of disjunct paleoplates and arc systems. The most important turnover occurred during the late Caradoc (figure 17.8B), near the base of the graptolite-based linearis Zone (TS.5c–d boundary). The taxa that originated during the early to mid Caradoc (TS.5a–b) were mainly cosmopolitan or widespread in the tropics and adjacent areas. In contrast, those originating during the late Caradoc (TS.5d) are endemic or show relatively high degrees

169

     . .           . tions within the Hirnantia and Edgewood faunas, respectively.

Pentamerides—First Pentamerides—Last

1c–1d

2b–2c

1c–1d

2b–2c

6a–6b

Number of Genera

170

6a–6b

FIGURE 17.8. Pentameride, atrypide, and rhynchonellide biodiversity. A, first and last appearances of the pentameride taxa measured in numbers of genera. B, standing diversities of the cyrtomatodont atrypides and rhynchonellides together with the advanced deltidiodont pentamerides.

of provinciality. During the late Caradoc and early Ashgill, rhynchonellides invaded the epicontinental seas of both Laurentia and Siberia, becoming the dominant components of the shallower-water brachiopod fauna. During much of the Ashgill (TS.6a–b) the rhynchonellides exhibited remarkably low rates of origin and high rates of extinction. New genera mainly appeared during the Hirnantian, when short-lived, possibly opportunist forms such as Plectothyrella and Thebesia developed surprisingly widespread distribu-

Atrypida (PC). The Atrypida originated during a late Llanvirn (Llandeilian) radiation (figure 17.8B). Throughout the history of the order its taxa were confined to the tropics, usually occupying subtidal carbonate shelf and ramp environments, commonly in shallow water (Copper 2001a). The first major radiation of the group occurred during the Caradoc and involved 13 genera. The Caradoc interval provided the opportunity for considerable experimentation within the order with the development of a variety of lophophore types and their supports. The origin of the group is still unclear, although smooth and slightly costate forms, Manespira and Anazyga respectively, may have occurred in the Llanvirn (Llandeilian). Nevertheless, by the later Caradoc to the mid Ashgill, anazygids such as Catazyga and Zygospira were very abundant on and around Laurentia, commonly forming shell pavements (Copper 2001b). Zygospira was most common in shallow-water facies, whereas Catazyga generally occupied deeper-water habitats. During the mid to late Ashgill, atrypid diversity declined; both Catazyga and Zygospira have yet to be reported from definitive Hirnantian rocks, although both subfamilies to which they belong reappeared during the Llandovery. In contrast, the spirigerinids diversified during the late Ashgill, replacing the early to mid Caradoc Pectenospira and Sulcatospira, both associated with carbonate mound facies in Kazakhstan. The spirigerinids increased in abundance and diversity in North America and Europe during the Hirnantian, displacing the incumbent anazygids (Copper and Gourvennec 1996). Septatrypinids such as Webbyspira and Idiospira, together with cyclospirids such as Rozmanospira and Cyclospira, evolved simultaneously during the Caradoc and Ashgill to complete the diversity of the group. The latter family probably originated in China or Kazakhstan and migrated westward during the late Caradoc (Popov et al. 1999a). Spiriferida (JR). Traditionally spiriferide origins were tracked back to the earliest Silurian. However, recent investigations suggest that the order originated during the Ashgill. Eospirifer from the middle Ashgill of East China is the earliest known representative of the order Spiriferida (Rong et al. 1994).

Brachiopods The major apomorphies of the spiriferides were established by the mid Ashgill, including a strophic shell with well-developed interarea and a brachidium of laterally directed spiralial cones (Carter et al. 1994). This is thus the last major taxonomic group of Brachiopoda to appear in the Ordovician. The taxon Eospirifer praecursor Rong et al. 1994 occupied a wide range of environments from BA 2 to BA 5. It occurred very abundantly in BA 2 and also occurred rarely as part of the Foliomena Fauna (BA 5). Athyridida (JR). To date, information is sparse on the origin and early evolution of the athyridides. Nevertheless, the earliest athyridides occur in Upper Ordovician rocks, represented by a few genera whose mutual relationships are not well known. The earliest known athyridide is Kellerella Nikitin et al. 1996, which appeared in Kazakhstan during the multidens Zone (TS.5b, early to mid Caradoc). At that time, the Kazakhstanian terranes were located as small microplates and island arcs adjacent to East Gondwana. Although the early Caradoc was a time of rapid morphological diversification in atrypids, a brachidium with spiral jugal processes had just appeared in the earliest known athyridids. They are impunctate and smooth (Alvarez et al. 1998) and include taxa lacking the septalium (Kellerella) and those possessing a welldeveloped septalium supported by a median septum (Nikolaispira). The fauna is associated with carbonate buildups and is characterized by a predominance of both Kellerella ditissima and Nikolaispira rasilis (Nikitin et al. 1996). The third genus of this major group, Hindella (= Cryptothyrella) is common in Hirnantian and Lower to Middle Llandovery rocks in the Northern Hemisphere. The earliest occurrence of Hindella may be in middle Ashgill rocks in Europe. However, a monospecific assemblage of Cryptothyrella? sp. occurs in strata, suggestive of BA 3–4, in the lower Ashgill of the Kerman and Tabas regions, east-central Iran (Bassett et al. 1999a). This early occurrence of the athyridides is significant; the group has not been reported in the coeval high-latitude faunas of West Gondwana. Fu (1982) established two genera, Apheathyris and Weibeia, from the Caradoc rocks of northwestern China. Apheathyris, despite reservations (Nikitin and Popov 1996), is retained within the athyridides. The systematic position of Weibeia is, however, less secure

because the characters of the brachidium are currently unknown and, moreover, Weibeia possesses a pair of long, parallel brachial plates, a feature atypical of the athyridides. Hyattidina and Whitfieldella occur in the higher Ordovician. The earliest species of Hyattidina is not well known, with Hyattidina? sulcata possibly the earliest record of this genus, from the Kiln Mudstones, Girvan, Scotland (Williams 1962). The earliest representatives of Whitfieldella are reported from the Hirnantia Fauna (TS.6c) in southwestern China (Rong 1979). Alvarez et al. (1998) summarized these data in their revised classification of the order. Based on these data, the athyridides probably originated in the regions of Kazakhstan, North China, and Qaidam, possibly in the early Caradoc (TS.5b). They may have been derived from an atrypide group by a substantial change of the direction of spiralia. ■

Ecologic Diversification (PMS, DATH)

The Ordovician Radiation marked the transition from the Cambrian EF to the Paleozoic EF (Sepkoski and Sheehan 1983; Sepkoski 1991a). Increasing diversity was accompanied by the rise to ecologic dominance of many new groups of organisms with the complexity of the ecosystem increasing markedly, especially in local communities. New adaptive strategies allowed some organisms to move into previously unoccupied habitats, and many more groups invaded habitats that had been occupied by fewer clades in the Cambrian EF. Compared with the many taxonomic studies of the Ordovician Radiation, relatively few studies have addressed ecologic changes among animals during the radiation, and this remains a fertile area of research (Droser and Sheehan 1997b). Ecologic changes can be grouped into autecologic changes, as new morphologies evolved that better adapted animals to the physical environment, and synecological changes, as animals adapted to increasing competition and predation. Competition for food may have been a driving force in the Ordovician Radiation; the success of brachiopods may have been due in great measure to their low metabolic rates (Rhodes and Thompson 1993), which allowed them to be successful in oceans of the Early Paleozoic, when food was in short supply

171

     . .          . (Bambach 1993; Martin 1998). Low food requirements were also characteristic of many other dominant early Palaeozoic groups, such as echinoderms, bryozoans, and sponges. Brachiopods have among the lowest metabolic rates of any marine organisms; the summary of Peck (2001:102) is worth repeating: “these (metabolic) attributes make them less able to compete for resources when they are in plentiful supply, but make them very strong competitors in a world where resources are limited or temporally restricted. They can survive where others starve from lack of resource supply, making them very resilient in difficult times.” Furthermore, brachiopods may have been well suited to exploit the Ordovician seas because their calcitic shells preadapted them to life in the calcite seas of the period (Stanley and Hardie 1999). The addition of new kinds of carnivores also increased community complexity; many species faced increased competition for food and at the same time were prey for increasingly efficient and specialized carnivores. These pressures must have driven adaptations in many directions, but the general trend was toward increased specializations and decreasing niche sizes. Communities from across marine environmental transects increased in diversity (Sepkoski 1988; Patzkowsky 1995c), with the exception of high-stress settings, such as nearshore environments, where organisms tended to be broadly niched and diversity changed little throughout the Phanerozoic. Bambach (1983) recognized 20 distinct megaguilds in the modern marine ecosystem; suspension feeders, herbivores, carnivores, and detritus feeders occurred across pelagic, infaunal, and epifaunal habitats that were further subdivided into mobile or sedentary lifestyles that were in turn subdivided into low or high tiers. Of the 20 megaguilds in the modern oceans, only 10 were occupied in the Cambrian EF, and this increased to 14 in the Paleozoic EF. Many new clades entered the various megaguilds during the Ordovician (figures 17.9, 17.10). The implications of differing community complexity and biodiversity are currently being examined in conservation biology. Increased complexity enhances the amount of resources captured and available to the community as a whole (Fridley 2001; Cardinale et al. 2002). This increase is driven especially by more efficient utilization of food and resources as organisms become more specialized for particular

45

Length (mm) of Longest Shell

172

40 35 30 25 20 15 10 5 0

Genera FIGURE 17.9. Increase in size of Plectorthoidea and Orthoidea brachiopods through the Ordovician radiation. The graph compares the largest individuals of genera illustrated in the Treatise (Williams and Harper 2000c) prior to the radiation (Upper Cambrian and Tremadocian, 25 genera, black columns) with those after the radiation (Caradoc and Ashgill, 53 genera, white columns). An increase in maximum size of brachiopods is indicated, although the method provides only a coarse estimate of the nature of this increase.

food-gathering strategies. Thus, there may have been a positive feedback in the Ordovician radiation that enhanced the amount of resources available to a community as diversity and complexity increased. Ecologic aspects of the Ordovician radiation of brachiopods may have been decoupled from their phylogenetic radiation (Droser et al. 1997; Bottjer et al. 2001). Clades may become major players in local communities by diversifying either before or after they increase in abundance in local communities. In addition, since there is a strong geographic component to the radiation (Miller and Mao 1998; Harper and Sandy 2001), it may not be possible to understand global patterns until many regions are examined (Droser and Sheehan 1997b). For example, bivalve mollusks diversified and became common members of communities in high latitudes during the Arenig before they appeared in North America (Babin 1995; chapter 20).

Food Resources and Brachiopod Size Ranges Phytoplankton-based food resources in the oceans increased significantly in the Mid Ordovician (Tappan and Loeblich 1973; Tappan 1986; Vermeij 1987; Bambach 1993), but the level of food resources was still far below modern levels. Many of the ecologic

Brachiopods

LLANVIRN

shallow

ˇ SÁRKA FORMATION

environment

Paterula Community

Orthid brachiopods Community Group Paterula Community Group

ARENIG

Nocturnellia Community

KLABAVA FORMATION

deep

Acrotreta Community

Acrotreta Community Group

Rafanoglossa Community

TREMADOC

Acrotreta Community

MÍLINA FORMATION

ˇ TRENICE FORMATION

Leptembolon Community Hyperobolus Community

Obolus Community Group Westonisca Community

FIGURE 17.10. Time environment diagram of the onshore-offshore pattern of the Ordovician Radiation in the Prague Basin (after Mergl 1999).

changes of the radiation may have been associated with increased food supply, and this may have been crucial for the expansion to proceed. Brachiopods increased in size during the radiation (figure 17.9), possibly facilitated by increasing food supplies, a trend documented by Bambach (1993) for other organisms. Similarly, the increasing abundance of brachiopods in shell beds after the radiation commenced (Li and Droser 1999) signaled increased numbers of brachiopods, perhaps again facilitated by increasing food. Orthide brachiopod clades that were present both before and after the radiation increased substantially in maximum size. Figure 17.9 compares the largest individuals of each genus of the Plectorthoidea and Orthoidea illustrated in the Treatise (Williams and Harper 2000c) from the Upper Cambrian and lowest Ordovician (Tremadocian) with those from the Caradoc and Ashgill. Although the method is imprecise, the trend is so clear that it is almost certainly robust. Plaesiomys from the Upper Ordovician of the Mackenzie Mountains of northwestern Canada reached a diameter of 60 mm. Other clades of brachiopods that originated during the radiation attained similar or larger sizes. In Laurentia, for example, the strophomenides Nasutimena, Tetraphalerella, and Oepikina from the epicontinental seas of Laurentia attained sizes of 60–80 mm (Jin 2001; Jin and Zhan 2001). Moreover, several members of the rhynchonelloid genera Hiscobeccus and Lepidocyclus attained lengths of 30–40 mm during the Ashgill. Larger brachiopods were not simply exceptions but were com-

mon, especially in open marine communities. In addition, Jin (2001) has shown that epicontinental-sea brachiopods were typically much larger than their congeneric counterparts on the shelf margins of Laurentia. The alternative hypothesis that this is simply a classic example of Cope’s Rule is difficult to test. Nevertheless, the gradual increase in size during clade history up to a maximum, predicted by the rule, is not typical of the Early Paleozoic history of the Orthida; Cambrian and Early Ordovician orthides had a rather constant size prior to a short interval of marked size increase during the radiation. Although food supply increased, it was still far below that available to the Modern EF (Bambach 1993); brachiopods, bryozoans, and echinoderms would still have maintained the advantages afforded by their low metabolic rates. Large size can have many selective advantages, such as the ability to occupy space and resist predation, but these advantages could be realized only if there were sufficient energy sources for the animals to succeed at these sizes.

Autecology As each of the new brachiopod clades evolved, a range of morphological innovations appeared. These innovations improved the adaptation of each species to the local environment. Ushatinskaya (2000) has provided a clear understanding of the autecology of Cambrian brachiopods. The linguliformean brachiopods, which were part of the Cambrian EF, had expanded to fill most of the lifestyles of later nonarticulates.

173

174

     . .          . The Cambrian articulated brachiopods, however, had not diversified morphologically and were primarily biconvex, epifaunal, and attached by a pedicle. Two of the most significant morphological changes during the Ordovician Radiation were the development of planar shells in the dominant lineages of post-Cambrian strophomenides and the evolution of cyrtomatodont teeth. Strophomenides adopted planar profiles in which opposite valves were closely opposed. When strophomenides abandoned the globose, biconvex shell, they abandoned a feeding strategy that enclosed the feeding currents and allowed an animal with very little biomass to occupy a relatively large space (Peck 2001). The planar shells had very little space for filter-feeding currents, and the strophomenides underwent a complete overhaul of the feeding method preferred by most brachiopods. The planar shells also permitted strophomenides to adopt a recumbent lifestyle (Leighton and Savarese 1996). This free-lying, reclining megaguild (Bambach 1983) is lacking in modern brachiopods, and as a result there is considerable uncertainty as to how these organisms functioned. Most Ordovician strophomenides probably lived concave upward on fine-grained substrates, with most utilizing “snowshoe” adaptations or, among geniculate forms, iceberg adaptations (Thayer 1975). Many other strophomenides lived on firm substrates, and they were often common in reefs. The cyrtomatodont teeth provided more substantial hinging of the shells, possibly partly as a means to inhibit predators and protect the animal from desiccation and the effects of turbulence. The mechanical hinge also lessened the expenditure of muscular energy needed to keep the shells aligned. Growth of the cyrtomatodont ball-and-socket articulation required new styles of shell formation including an emphasis on resorption of shell material to allow the socket to increase in size during growth. The cyrtomatodont teeth appeared first in rhynchonellides during the Darriwilian. Atrypides followed, then athyridides in the early Caradoc and spiriferides in the Ashgill. Many morphological changes were adaptations that had less extreme implications for brachiopod lifestyles. Some species developed adaptations that served multiple functions; for example, the evolution of plications served a dual purpose of preventing large particles from entering the chamber while at the same

time increasing the strength of the shell. Evolution of “snowshoe morphologies” (Thayer 1975) allowed brachiopods to inhabit soft substrates. These are two examples of the morphological innovations that permitted brachiopods to compete with other benthos for resources, defend against predation, and interact with the physical environment. Each of the radiating clades had many unique morphological innovations, many of which have been widely discussed (Rudwick 1970; Alexander 2001; Leighton 2001). However, the sequence of appearance of innovations during the radiation has not been addressed. For example, do morphological innovations appear before, after, or during increases in community complexity, as evidenced by increasing numbers of species and variety of guilds in a given habitat?

Synecology The environment, which was driving evolution, included both physical and biotic factors. There were probably modest changes in the physical environment during the radiation, and some brachiopods moved into niches that previously had not been occupied by brachiopods. However, the biotic environment was in the throes of a massive reorganization and expansion. As organisms diversified and autecologic changes accumulated, each species in the local communities was faced with a new synecological setting with increased competition for resources and increased predation. The relative simplicity of Cambrian EF communities contrasts with the ecologic diversity of those following the Ordovician Radiation (see Cocks and McKerrow, figures 2 and 11, in McKerrow 1978). The Ordovician Radiation set the style of communities for the remainder of the Paleozoic. The similarity of ecologic style was recognized by Walker and Laporte (1970) and later formalized by Sepkoski (1981a, 1991b) as the Paleozoic EF. Boucot (1975, 1983) divided the Paleozoic EF into a series of Ecologic Evolutionary Units during which communities were stable in the sense that diversity and ecologic structure were maintained. Only occasionally did groups move into new Benthic Assemblages (BAs). Specieslevel diversity and the various morphological types present in the community remained constant for tens of millions of years (e.g., Watkins et al. 2000).

Brachiopods By the end of the Ordovician Radiation brachiopods were established as important, often dominant members of communities in open marine settings. The communities were arrayed along environmental gradients (Springer and Bambach 1985), conveniently grouped into a series of six progressively deeper BAs (Boucot 1975). Although arranged from shallow (BA 1) to deep (BA 6), the assemblages did not have a consistent depth from region to region or sharp boundaries; rather, the BAs were controlled by parameters of the physical environment, including decreasing wave energy with depth, depth of penetration of sunlight, temperature gradients, nutrient supply, and oxygen levels. This pattern, which developed during the radiation, was maintained through the remainder of the Paleozoic (Boucot 1983). Changes in benthic communities from the Cambrian EF to the Paleozoic EF are readily apparent. However, limited knowledge about the relative timing of changes makes it impossible to choose between two hypotheses: (1) the radiation involved displacement of the Cambrian EF by members of a more competitive Paleozoic EF, or (2) a physical disruption of the Cambrian EF allowed radiation of the Paleozoic EF despite the advantages provided by incumbency to the Cambrian EF (Sepkoski 1991b). In fact, there is still uncertainty as to whether the Cambrian EF actually declined or if it continued in preradiation abundance and diversity but was simply less conspicuous because of a vastly more common and more diverse Paleozoic EF (Westrop et al. 1995). Synecology of the Rhynchonelliformean Brachiopods Increasing taxonomic diversity by itself is a major component of the radiation. But taxonomic diversity, for example, the number of families within orders, also provides a proxy for ecologic diversity because species within families commonly share many morphological features that determine their lifestyles. The enormous increase in numbers of orders, together with the numbers of families and genera (figure 17.1), indicates an enormous increase in the different lifestyles of rhynchonelliformean brachiopods, which were incorporated into the marine ecosystem during the Ordovician Radiation. The linguliformean fauna developed as a holdover from the Cambrian EF, while the craniiformeans played a relatively obscure role in

the radiation as a whole (see Popov et al. earlier in this chapter). The rhynchonelliformeans, however, dominated the event. Distribution of modern articulated brachiopods is entirely by nonplanktotrophic reproduction, producing relatively brief larval stages. Groups with nonplanktotrophic larvae have smaller numbers (lower fecundity) than those of planktotrophic groups. Nonplanktotrophic larvae and low fecundity probably also were typical of Ordovician orthide brachiopods, based on pallial markings (Law and Thayer 1990). However, Brunton and Cocks (1996) noted that some strophomenides had developed a small shell (protegulal node) when their larvae settled, implying a prolonged larval feeding stage such as that in modern nonarticulated brachiopods. In terms of a radiation, the brief larval stages of nonplanktotrophic groups should have resulted in more narrow distributions than planktotrophic groups. Nonplanktotrophs might speciate more readily than planktotrophs because of their more restricted geographic ranges; this could have promoted rapid radiation ( Jablonski 1986). A comparison of the rates of radiation of planktotrophic and nonplanktotrophic groups (both brachiopods and other phyla) in a single region during the Ordovician could provide a test for these hypotheses. Although nonplanktotrophic larval stages may have aided brachiopods during the radiation, it may also have been part of their ultimate decline. Valentine and Jablonski (1983) suggested that the decline of post-Paleozoic brachiopods might have been related to developmental inflexibility, which resulted in their replacement by bivalves, with a greater developmental flexibility. Open Marine Environments More than 50 morphologically distinct ecologic guilds have been identified in open marine communities of the Paleozoic EF (e.g., Bambach 1983; Watkins 1991, 1993). Many more guilds have yet to be recognized. After the radiation most brachiopods belonged to one of seven guilds (Bambach 1983; Watkins 1991, 1993): (1) cap-shaped “inarticulates,” (2) infaunal “inarticulates,” (3) strongly biconvex pedunculate, (4) strongly biconvex alate pedunculate, (5) planoconvex to weakly biconvex pedunculate, (6) relatively flat free-lying, and (7) inflated free-lying articulates.

175

176

     . .          . There were also many minor guilds, such as cemented forms, brachiopods that attached to echinoderms in upper tiers and inhabitants of cryptic spaces. Additional brachiopod guilds, such as spinose, semiinfaunal productids, appeared later in the Paleozoic. The guilds presumably parceled ecospace so that their members were in greater competition among themselves than with members of other guilds. Not all guilds are present in any given local community, but many guilds have several species. Prior to the radiation, orthides and pentamerides were the most common articulated brachiopods, with billingsellids less common. The primary rhynchonelliformean brachiopod orders that diversified during the Ordovician Radiation (the Orthida and Strophomenida) had contrasting lifestyles. Among the orthides only the orthidines were prominent prior to the radiation, and they maintained and even increased in diversity during the radiation. The dalmanellidines radiated substantially, becoming as diverse as the orthidines. Other radiating groups included the Atrypida and Rhynchonellida (figure 17.8), Orthotetida, and Athyridida. The Billingsellida (sensu stricto) expanded during the early phases of the radiation but then declined through the Ordovician. The Pentamerida declined from preradiation levels but staged a comeback in the Silurian. The Spiriferida appeared later in the Ordovician and radiated in the Silurian. Each of these newly radiating groups had evolved distinct morphological characteristics that allowed them to exploit the environment in different ways. A result of this radiation was that the number of brachiopod guilds in communities increased substantially. Marginal Environments Extreme or marginal environments were occupied by low-diversity communities. Shallow, restricted environments with soft substrates often were colonized by the Lingula Community type characterizing BA 1 of Boucot (1983). Shallow-water environments subjected to heavy wave action typically have low diversity faunas of broadly niched species capable of withstanding unpredictable environmental changes such as desiccation, temperature variations, variations in normal waves, and occasional storms. Diversity has remained low in these environments throughout the

Phanerozoic (Bambach 1983), probably because of the need to maintain relatively broad niches and tolerances to live successfully in unpredictable settings. At the other end of the marine spectrum, BA 6 communities lived in dysaerobic, relatively deep offshore environments that were transitional between open marine assemblages and the graptolite shale environments that commonly had insufficient oxygen for benthic animals. An example of this is the Foliomena Community type (Sheehan 1973; Cocks and Rong 1988; Harper and Rong 1995; Rong et al. 1999), with small, thin-shelled brachiopods commonly with soft-substrate morphologies that were associated with vagile trilobites. Predation Predation is an important aspect of synecological interaction. The effects can be highly varied, in some cases excluding groups from a region and in other cases culling dominants and thus providing opportunities for less competitive groups to survive. Crushed brachiopod shells are not uncommon following the Ordovician Radiation, and many examples of repair of damaged shells have been described, indicating that predation was dominated by durophagous predators (arthropods and possibly nautiloids and asteroids), together with fungal borers (Alexander 1986). Drilling, if present, was not common until the Devonian, when fishes and possibly gastropods became important predators (Leighton 2001). A variety of morphological changes may have been adaptations against predators. Examples include increasing shell strength with a fold and sulcus, plications, cyrtomatodont tooth structures, protective structures such as spines, coverings of the pedicle opening, punctae, and chemical defenses that made the animals inedible. Pace of Synecological Change The transition from communities dominated by the Cambrian EF to communities dominated by the Paleozoic EF moved progressively from shallow to deep water across the shelf and upper slope (Sepkoski and Sheehan 1983; Sepkoski and Miller 1985). The progressive dominance of brachiopods in more offshore environments has been independently confirmed by studies of the Ordovician brachiopods of the Anglo-

Welsh region (Lockley 1993) and marginal Laurentia (Patzkowsky 1995b). The pace of onshore to offshore synecological change seems to have been moderately rapid. Time-environment diagrams generally show that the change from Cambrian EF communities to Paleozoic EF communities across shelf environments began in the Arenig and was completed during the Caradoc in North America (Sepkoski and Sheehan 1983: figure 7). It may have been more rapid in Bohemia, where it was completed in the Llanvirn (Mergl 1999; and see figure 17.10). In Argentina the change took place in the early Arenig, possibly during a short interval at the end of the early Arenig (Waisfeld et al. 1999). Few synecological field studies have actually examined the Ordovician Radiation. Li and Droser (1999) showed that brachiopod abundance increased rapidly near the base of the Whiterockian stage in North America. Finnegan and Droser (2001) found that in western North America, although brachiopod abundance increased rapidly, the diversity of level-bottom communities lagged during the early Whiterockian. Brachiopod-dominated communities were characterized by relatively low diversity despite the presence of very abundant brachiopods. Furthermore, few of the new clades that appeared during the radiation were present in the initial communities. By the end of the Whiterockian, community diversity of brachiopods had increased substantially, and the new clades were present.

Paleoecologic Levels Comparing ecologic changes at different times in the geologic record can be accomplished only if there is a standardized method of assessing changes. A series of four standardized paleoecologic levels of change has been developed (Droser et al. 1997, 2000; Bottjer et al. 2001). These levels are ordered from those with the farthest-reaching ecologic change to those that primarily involve local communities. First-level changes represent the advent of a new ecosystem. The colonization of land was proceeding during the Ordovician, but this first-level change did not involve brachiopods. Second-level changes involve appearance or changes in the ecologic dominants in the ecosystem. The change from trilobitedominated level-bottom communities of the Cam-

Number of Genera

Brachiopods

1c–1d

2b–2c

6a–6b

FIGURE 17.11. Standing diversities of all major rhynchonelliformean groups compared; measured as numbers of genera.

brian EF to domination by brachiopods, echinoderms, bryozoans, and other groups of the Paleozoic EF (figures 17.10, 17.11) was a second-level change. The change in shell morphology and feeding methods of the strophomenides was a second-level change that allowed them to move into ecospace previously unoccupied by brachiopods. Increasing levels of predation and the way brachiopods responded to predation are also second-level changes that significantly restructured communities and interactions at the local level. Much of the story of the radiation of brachiopods took place at the third level, when megaguilds were filled to Paleozoic EF levels (Bottjer et al. 2001). A third-level increase in epifaunal tiering, largely produced by echinoderms (Ausich and Bottjer 1982), may have affected brachiopods, since food was removed from the water column prior to reaching base-level brachiopods. Fourth-level changes involved formation of new paleocommunities, diversification of the newly evolved clades, and increase in maximum size of brachiopods. ■

Conclusions

The Ordovician radiation of the brachiopods was dominated by the expansion of rhynchonelliformean

177

178

     . .          . clades (figure 17.11), the linguliformeans and craniiformeans playing more minor but nevertheless significant roles in the event. The linguliformeans were already an important part of the Cambrian fauna but generally migrated into deep-water habitats during the Ordovician, when the acrotretides became the dominant group. The craniiformeans had a limited migrational capacity and apparently moved diachronously across the various Ordovician paleoplates and arc systems. Following a pre-Caradoc history associated with Baltica and various terranes located between Baltica and East Gondwana, the group diversified across a range of latitudes. The two main groups of deltidiodont rhynchonelliformeans, the biconvex and pedunculate orthides together with the plano- to concavo-convex recumbent strophomenides, participated in a series of stepwise radiations. Mid Ordovician diversification involved both the displacement and expansion of many clades into deep-water habitats (Harper et al. 1999a, 1999b). Among the cyrtomatodonts, the earliest occurrences of the orders Athyridida, Atrypida, and Spiriferida were concentrated in low-latitudinal areas between the Baltica and East Gondwana, that is, somewhere between equatorial Gondwana and the Baltica, during the Caradoc (see, e.g., Jaanusson 1979; Popov et al. 1997). This region was occupied by a number of peri-Gondwanan plates and volcanic island arcs (such as North and South China together with the Kazakhstanian terranes), providing a locus for the origin and initial radiation of the earliest spire-bearing brachipods. The rhynchonellides, however, may have origi-

nated on Laurentia, radiating during the Mid and Late Ordovician across a spectrum of shallow-water carbonate environments. The complex taxonomic aspects of the radiation are clear signals of an equally complicated ecologic event involving the transition from the Cambrian to Paleozoic evolutionary faunas, the availability of nutrients, the diversification of predators, and life in a calcitic sea. The hypothesis that the expansion of clades prompted the occupation of ecospace by both alpha (within community) and beta (between community) diversity (Sepkoski 1988) requires rigorous investigation. Nevertheless, during the event the number of megaguilds increased while a number of new clades entered the megaguild structure. This development of a range of new adaptive strategies among the Brachiopoda markedly changed the Paleozoic seascape.

 Harper thanks the Danish Natural Science Research Council (SNF) for support and Anne Haastrup Hansen (Geological Museum, Copenhagen) for redrafting most of the figures. Research was also supported by grants (EAR-9706736, EAR-9910198) to Sheehan from the U.S. National Science Foundation. Popov and Bassett acknowledge support from the Royal Society and the National Museum of Wales. The chapter was improved by constructive and thoughtful reviews from Mark Patzkowsky (Pennsylvania) and Tony (A. D.) Wright (Leicester) together with editorial comments from Barry Webby.

18

Polyplacophoran and Symmetrical Univalve Mollusks Lesley Cherns, David M. Rohr, and Jirˇ í Fry´da

G

eneralized taxonomic relationships and biodiversity patterns of the Ordovician representatives of two small molluscan groups—the scleritebearing polyplacophorans (or chitons), and the symmetrical univalves (Tryblidiida and Bellerophontida)—are discussed here.

■

Polyplacophorans (LC)

The polyplacophorans (or chitons) have a long fossil record (Cambrian–Recent) but are rare fossils and are mostly known from isolated intermediate sclerites. Head and tail plates, which commonly differ markedly in morphology, size, and ornament from the series of intermediate plates, are notably few in the fossil record. A similar complement of eight sclerites to Recent polyplacophorans is apparent from rare articulated specimens (e.g., Rolfe 1981) and also a comparable paleoecology for most in shallow marine facies (e.g., Smith and Toomey 1964; Runnegar et al. 1979; Cherns 1999). The generally poor preservation potential of rocky shore paleoenvironments probably contributes to the sparse fossil record. The pattern of Ordovician chiton diversification is distorted by detailed records limited to only a few faunas. Those major faunas come from the Lower Ordovician of the United States and Australia (e.g., Bergenhayn 1960; Smith and Toomey 1964; Runnegar et al. 1979; Stinchcomb and Darrough 1995).

Other records are fairly sparse and isolated but include significant Upper Ordovician genera from Scotland (e.g., Rolfe 1981). The phosphatic, debatably polyplacophoran Cobcrephora from Australia (Bischoff 1981), from an allochthonous block of probable Late Ordovician age within the Lower Devonian Cuga Burga Volcanics (I. G. Percival pers. comm.), remains a problematic taxon. The higher taxonomy of polyplacophorans is far from settled. In the current scheme, almost all Early Paleozoic chitons are included in the subclass Paleoloricata (Cambrian–Upper Cretaceous), whose members lack an articulamentum and hence sutural laminae and insertion plates (Smith and Hoare 1987). In the subclass Neoloricata (Carboniferous–Recent), which includes all Recent chitons, only some Carboniferous lepidopleurine families lack insertion plates. Carboniferous lepidopleurid species previously assigned to Helminthochiton were reassigned to Gryphochiton because the Silurian type species H. griffithi and also Middle Ordovician H.? aequivoca apparently lack both sutural laminae and insertion plates (Smith and Hoare 1987). Hoare (2000: figure 5) proposed the paleoloricate family Helminthochitonidae, including Early to Mid Paleozoic Helminthochiton species, as a possible stem group for later neoloricates. Discussion of the early history and origin of polyplacophorans centers around genera of Upper 179

180

   . Cambrian–Lower Ordovician multiplated organisms that are questionably chitons, such as Matthevia, Hemithecella, and Preacanthochiton. Lower Cambrian microgenera from China (Yu 1987) were rejected as polyplacophorans by Qian and Bengtson (1989; also Sirenko 1997; Hoare 2000). Separate higher-level molluscan taxa were proposed for Matthevia (Yochelson 1966) and Hemithecella (Stinchcomb and Darrough 1995). Stinchcomb and Darrough (1995) established several new genera for elongate sclerites associated with Hemithecella but wished to exclude all these and also Preacanthochiton (Bergenhayn 1960) from the Polyplacophora because of what they recognized as “monoplacophoran-like multiple muscle scars” (Stinchcomb and Darrough 1995:52) and some valve asymmetry. However, Runnegar et al. (1979) argued that Upper Cambrian Matthevia represented the earliest known polyplacophoran and that its tall conical valves became replaced by flatter, elongate, and overlapping valves through younger Hemithecella to Chelodes. Matthevia, Hemithecella, and Preacanthochiton are accepted here as paleoloricates (Runnegar et al. 1979; Smith and Hoare 1987; Hoare 2000), although Sirenko (1997) proposed their exclusion from his classification into two orders and eight families. Hoare’s (2000) modified classification (from Sirenko 1997), including all those genera, has only four families present in the Ordovician. Paleozoic chiton phylogeny still has much to resolve, and the recent description of a plated aplacophoran from the Silurian (Sutton et al. 2001) has further widened the debate. The data in figure 18.1 include only published species and follow Hoare’s (2000) scheme except that the Septemchitonidae is restored. The elongate, robust sclerites of Tremadocian Hemithecella, Conodia, Robustum, and Calceochiton, together with Chelodes and Eochelodes, are in the broad grouping of the Mattheviidae (figure 18.1A–F). Chelodes is long-ranging and occurs widely in Ordovician and later Silurian rocks. Tremadocian Preacanthochiton (figure 18.1G), which has small conical plates with a pustular ornament, in the Preacanthochitonidae, was placed by Runnegar et al. (1979) outside its Matthevia-Chelodes lineage. In the Gotlandochitonidae, sclerites are wider than long, with weak to clearly defined shell areas. Gotlandochiton, Paleochiton, and Ivoechiton (figure 18.1H–J) occur in an important chiton assemblage from the early Arenig

(Smith and Toomey 1964). The Helminthochitonidae (Sirenko 1997; figure 18.1K, L) includes Kindbladochiton from that same Arenig assemblage (Smith and Toomey 1964) and Helminthochiton?. The Septemchitonidae includes Upper Ordovician Septemchiton and Solenocaris (figure 18.1M, N), which have very elongate, narrow sclerites. Priscochiton (figure 18.1O), based on very limited material, lacks family assignment. The pattern of Ordovician diversification in figure 18.1 shows that the Early Ordovician was an important period of polyplacophoran radiation, particularly on the low-latitude Laurentian margin. Upper Cambrian–lower Tremadocian assemblages of Matthevia, Hemithecella, Preacanthochiton, Conodia, and Robustum occur in stromatolitic, shallow marine carbonates and dolomites (Yochelson 1966; Runnegar et al. 1979; Stinchcomb and Darrough 1995). Similar paleoenvironments were associated with Smith and Toomey’s (1964) diverse lower Arenig Laurentian assemblage and on the Gondwanan margin for Chelodes whitehousei from Australia (Runnegar et al. 1979). The Lower Ordovician shows the highest diversity of chitons with 10 genera and 20 species. This peak of diversity compares with the initial radiation of bivalves, except that chitons were diversifying on the low-latitude Laurentian carbonate margin, whereas bivalves were confined to higher latitudes and siliciclastic-dominated settings on Gondwana and Avalonia (e.g., chapter 20). The data from the Middle Ordovician are currently restricted to very few, isolated records. Chelodes? mirabilis from Alabama represents Laurentia, and Helminthochiton? aequivoca from Bohemia represents a peri-Gondwanan terrane, Perunica (chapter 5). The Upper Ordovician shows more diversity, with five genera and nine species. Solenocaris includes the first record from Baltica. Laurentian chitons are represented by Solenocaris, Chelodes, and Septemchiton from the Midland Valley of Scotland, the last two also from the United States. Eochelodes from Bohemia represents Perunica. Isolated and localized occurrences preclude useful paleoenvironmental interpretation. There are no records of polyplacophorans from late Ashgill rocks. The Silurian record of polyplacophorans shows diverse assemblages particularly from shallow carbonates on Gotland (Bergenhayn 1955; Cherns 1998a,

Polyplacophoran and Symmetrical Univalve Mollusks

UPPER

TIME SLICES

North American British Ashgill

Cincinnatian

ORDOVICIAN

443 Ma

6c 6b 6a 5d

Mohawkian Caradoc

GLOBAL SERIES & STAGES

Reg. series

5c 5b 5a

Whiterockian Llanvirn

DARRIWILIAN

MIDDLE ORDOVICIAN

460.5 Ma

4b 4a 3b 3a 2c 2b 2a

Ibexian

1d Tremadoc

LOWER

TREMADOCIAN

ORDOVICIAN

Arenig

472 Ma

4c

FIGURE 18.1. Range chart for genera of Ordovician polyplacophorans. Information on ranges is given as accurately as possible, for most genera to one time slice (TS ), and ranges are shown to time-slice boundaries. Numbers against each line indicate species recognized for Lower, Middle, and Upper Ordovician parts of ranges, as appropriate. Dotted lines signify gaps within ranges through which a genus can be assumed to continue. Family assignments: A–F, Mattheviidae; G, Preacanthochitonidae; H–J, Gotlandochitonidae, K–L, Helminthochitonidae; M–N, Septemchitonidae; O, indeterminate family.

1c 1b 1a

490 Ma

1998b). Though still rare, these chiton-rich assemblages emphasize the bias thrown on analyses of diversity where the fossil record otherwise is very sparse. ■

Tryblidiids and Bellerophontids (DMR, JF)

Paleozoic symmetrical univalved mollusks (Tryblidiida and Bellerophontida) are one of the most discussed groups of the phylum Mollusca. Correct evaluation of phylogenetic relations of these ancient mollusks to other molluscan groups is crucial for our understanding of the evolution of all molluscan

classes of the subphylum Cyrtosoma. Opinions on their phylogenetic relations were recently analyzed in several papers (e.g., Starobogatov 1970; Peel 1991a, 1999b; Wahlman 1992; Geyer 1994). However, these analyses resulted in quite different phylogenetic models and caused the erection of many new orderand class-level names. In addition, their subsequent usage and generic content have often been very different, and they have disregarded their original diagnoses (e.g., Monoplacophora and Tergomya). Tryblidiida with cap-shaped shells and isostrophically coiled Bellerophontida belong to the largest groups among the Ordovician symmetrical univalved mollusks. In

181

182

   . addition, several minor molluscan groups such as Helcionelloida (= Eomonoplacophora), Cyrtolitida, and Pelagiellida are known also from Ordovician strata and probably played a significant role in early molluscan phylogeny (see Starobogatov 1970; Geyer 1994; and Gubanov and Peel 2000 for discussion). Because of the absence of a generally accepted classification of the Paleozoic symmetrical univalved mollusks, we analyzed only the two largest Ordovician groups, Tryblidiida and Bellerophontida, which can be easily distinguished by their shell morphologies. We (DMR, JF) feel confident that both groups are probably polyphyletic as suggested by many authors (see later in this chapter and chapter 19). Nevertheless, their more detailed analysis is impossible without future reevaluation of all available data as well as reconsideration of the significance of some characters (e.g., muscle scars). The main goal here is to illustrate the Ordovician biodiversity patterns of Tryblidiida and Bellerophontida and to point out some problems in the evaluation of their phylogenetic relationships. Refer to the next chapter in this volume on the Gastropoda for the definitions of diversity parameters.

Tryblidiida (= “Monoplacophora”) The Ordovician groups placed in Tryblidiida by Geyer (1994) as well as Achinacelloidea are united here in one group. The systematic position of the latter group is still uncertain, and it has been considered to belong to the class Gastropoda (e.g., Knight and Yochelson 1958; Peel and Horny´ 1999) or to the “Monoplacophora” (Horny´ 1963; Runnegar and Jell 1976). At present there is no reliable evidence about whether members of both these groups, Tryblidiida and Achinacelloidea, are torted or untorted. As correctly noted by Harper and Rollins (1982, 2000) and Wanninger et al. (2000), the usage of muscle scar patterns as a key for determination of torsion in fossil mollusks has no basis in fact. In addition, the belief that symmetrically arranged multiple muscle scars are evidence for symmetrical, segmented anatomic arrangement of the shell body is false. These two speculations were evoked by discoveries of living “monoplacophorans” (Neopilinidae), which seem to confirm a theoretical concept of a stem group for the conchiferan mollusks as expressed by Odhner (in

Wenz 1940). These facts also opened a controversial debate over which characters of neopilinid anatomy are derived and which are primary (see Haszprunar and Schaefer 1997 and references therein). Anatomy of the modern Neopilinidae seems to be a result of ancient molluscan organization as well as specialization. Interpretation of the body plan in the Paleozoic Tryblidioidae as being closely similar to that of the living Neopilinidae is thus problematic. Moreover, the integration of both groups into one natural group is uncertain because there is no reliable evidence that they are really closely related. As shown earlier in this chapter, we are very far from being able to determine natural groups within the Ordovician mollusks with cap-shaped shells. The Ordovician Tryblidiida are known from equatorial as well as high-latitude regions. The majority of genera were described from the paleotropical realm (Laurentia, Baltica, and Siberia), and they had very limited geographic ranges. The quantitative diversity pattern of the Ordovician Tryblidiida cannot be described in detail because of a lack of detailed stratigraphic data for the majority (mainly Siberia and Baltica). Their genera are often based on limited material (sometimes even on a single shell with well-preserved muscle scars). The Ordovician Tryblidiida reached their highest diversity during the Early Ordovician (18 genera), and most genera were restricted to this time interval. During the Mid and Late Ordovician their diversity is slightly lower (14 and 13 genera, respectively). Only about one-quarter of the genera crossed the Ordovician/ Silurian boundary.

Bellerophontida (Amphigastropoda) The class-level assignment of bellerophontoidean mollusks has often been discussed during the past 50 years. Considerations about the morphology of soft parts of these coiled, bilaterally symmetrical mollusks divided paleontologists into several groups. The question essentially resolves to whether all or part or none of the bellerophontiform mollusks were untorted, exogastrically oriented monoplacophorans or torted, endogastrically oriented gastropods (see Yochelson 1967; Harper and Rollins 1982, 2000; Peel 1991a, 1991b; Wahlman 1992; Fry´da 1999b, for review). Recently Fry´da (1999b) described a multiwhorled

Polyplacophoran and Symmetrical Univalve Mollusks

Amphigastropoda dtot 3 dnorm di 0

dtot

oi e 0 i

dnorm

1

di

1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c TREMADOCIAN

LOWER ORDOVICIAN

489 Ma

DARRIWILIAN

M ORDOVICIAN

472

UPPER ORDOVICIAN

460.5

443

FIGURE 18.2. Plot illustrating generic biodiversity and turnover rates of Amphigastropoda (bellerophontiform mollusks). Lines show the biodiversity (dtot = total diversity; dnorm = normalized biodiversity, and di = number of genera per m.y.). Bars show turnover rates (white = rate of originations per m.y. [oi ], black = rate of extinctions per m.y. [ei ]). For definitions of the parameters used, see chapter 19.

protoconch in the type genus Bellerophon and interpreted it as a true larval shell. This fact suggests that this genus (and thus, also the Bellerophontoidea, Amphigastropoda) does not belong to the Archaeogastropoda (Fry´da 1999c, 2001). On the other hand, some new data on the protoconch morphology indicate that the Paleozoic bellerophontiform mollusks represent a polyphyletic group (Fry´da 1999c). Recently Harper and Rollins (2000) concluded that the bellerophontoideans and the coiled and high-domed “monoplacophorans” were gastropods. The Early Paleozoic bellerophontoidean mollusks were a very successful group and occupied a wide range of depositional environments (Peel 1977, 1978; Wahlman 1992). In his excellent monograph, Wahlman (1992) summarized knowledge of the paleoecology of Ordovician bellerophontoidean mollusks and clearly showed that they led a variety of modes of

life. This fact probably explains their cosmopolitan distribution. The oldest bellerophontoidean mollusks (family Sinuitidae) are known from Upper Cambrian strata. Their generic diversity was relatively low until the early Mid Ordovician. From approximately the beginning of the Darriwilian (TS.3b and 4a; see time slices in chapter 2), the number of genera per million years (m.y.) (di ) continually increased to the end of the Ordovician (figure 18.2). The mid Late Ordovician (TS.5c–6a) was the time of the highest diversity of the Ordovician bellerophontoidean mollusks (dnorm close to 14). The rate of originations per m.y. (oi ) was variable through the Ordovician with distinct peaks at the early Tremadocian (TS.1a), early Mid Ordovician (TS.3b), and early (TS.5a) and mid (TS.5c) Late Ordovician (figure 18.2). The rate of extinctions per m.y. (ei ) reached the highest values at the beginning of Arenig (TS.2a) and during the Ashgill (TS.6a and 6b). At present it is difficult to make a more detailed analysis of the turnover rates because the bellerophontoidean mollusks probably represent a polyphyletic group (see earlier in this chapter). It is concluded that the Ordovician symmetrical univalved mollusks form a relatively large group (about 75 genera), even though the majority of them are imperfectly known. They include several morphological groups (see Wahlman 1992 and Geyer 1994 for review) for which neither the phylogenetic relationships nor their relationships to extant molluscan groups are known. Biodiversity of the Tryblidiida seems to have decreased slightly during the Ordovician, and it was strongly affected by the extinction event close to the Ordovician/Silurian boundary. On the other hand, the number of the genera belonging to the Bellerophontida increased through the Ordovician. Their diversity pattern shows similarities with some gastropod groups (see chapter 19).

183

19

Gastropods Jirˇ í Fry´da and David M. Rohr

G

astropods are one of the most diverse groups of animals and include more than 100,000 living species. The number of fossil species must have been much higher, and estimates of gastropod diversity suggest about 13,000 extant and fossil genera (Bieler 1992). Gastropods’ rich and long-ranging fossil record (more than 500 million years [m.y.]), coupled with their occurrence in almost all marine, freshwater, and terrestrial environments, makes them a unique animal group for evolutionary, ecologic, and biogeographic investigations. Even the early evolution of the class Gastropoda seems to be well documented by rich fossil material from Ordovician strata. However, attempting a more detailed analysis of Paleozoic gastropods has revealed quite a serious problem, the core of which lies in the limited compatibility between neontological and paleontological taxonomic systems of the class Gastropoda. In contrast to the classification of living gastropods, that of fossil gastropods is based on only a limited number of characters, observable in their fossilized hard body parts. Correct determination of phylogenetic relationships in fossil gastropods decreases considerably with increase in their geologic age because of their uncertain relationships to modern gastropod groups and the frequent homoplastic similarity of their shells. Thus, it is quite evident that paleontologists, having only the possibility to evaluate a limited number of shell features, have very often worked with artificial groups, which in

184

reality do not illustrate the natural phylogenetic relationships among fossil gastropods. Higher classification of Ordovician gastropods has traditionally been based only on evaluation of their teleoconch characters (Wenz 1938–1944; Knight et al. 1960; Pchelintsev and Korobkov 1960). Certainly the teleoconch characters bear some phylogenetic signals, but because of frequent homoplasy, these features cannot produce a reliable evaluation of the higher taxonomic position of the fossil gastropods. Modern evaluation of the gastropod fossil record is also beginning to use high-level taxonomically significant characters such as the pattern of the gastropods’ early ontogeny, which has been reflected in their protoconch morphology. However, some recent discoveries such as the presence of openly coiled protoconchs in some Paleozoic gastropod groups (Hynda 1986; Dzik 1994c; Fry´da and Manda 1997; Bandel and Fry´da 1998, 1999; Fry´da 1999a, 1999b; and Nützel et al. 2000 and references therein) seem to complicate understanding of gastropods’ early evolution. Another very serious problem is the lack of modern data about their stratigraphic ranges. Many large Ordovician gastropod faunas have not been studied, or they rely on descriptions that were published about 100 years ago. These facts considerably limit analysis of the biodiversity of Ordovician gastropods. In this chapter, we analyze their generic diversity and discuss definition, phylogenetic relationships,

Gastropods and paleobiogeography of the major groups of Ordovician gastropods in order to illustrate the diversity patterns at higher taxonomic levels. ■

Diversity Measures

As shown in detail later in this chapter, there are many unsolved problems in the higher taxonomy of Ordovician gastropods. Classifications inferred only from teleoconch characters do not fit well together. For example, compare the commonly used alpha taxonomic approach of Knight et al. (1960) and Pchelintsev and Korobkov (1960) with results of the phylogenetic analyses of Wagner (1995a, 1995a). In addition, there is considerable inconsistency between the classifications inferred from teleoconch and protoconch characters (Fry´da 1999b). Because of the lack of a generally accepted classification of Ordovician gastropods, a relatively simple method is chosen for the description of the diversity patterns. The analysis is based on generic taxa, and only generally accepted higher taxa are used. Modifications of generic content of some higher taxa used for our analysis are mentioned in the text. There are many possible ways to measure biodiversity of fossil animals. Selection of appropriate parameters for quantification of their diversity depends on the longevity of the taxa relative to the duration of the time units used (e.g., Archibald 1993; Cooper 1999c). The present analysis of Ordovician gastropod generic diversity is based on more than 2,500 records of Ordovician gastropods belonging to about 140 genera. Nineteen time slices (TS) through the Ordovician are used in order to compare our results directly with those of other clades, even though this seems to be too detailed for generic analysis. In general, most Ordovician gastropod genera have ranges much longer (average = 14 m.y.; median = 12 m.y.) than the 1.5– 3 m.y. durations of the time slices used. Only about 6 percent of the Ordovician gastropod genera had ranges equal to or shorter than those of the time slices. For these reasons the following parameters were used for the analysis of biodiversity: (a) total diversity (dtot ), defined as the total number of genera that are recorded from the time slice; (b) normalized diversity (dnorm ), defined as the number of gastropod genera ranging through the time slice plus half the number of genera confined to the slice or ranging

beyond the time slice but originating or ending within it; and (c) genera per m.y. (di ), defined as the number of gastropod genera present within the time slice divided by its duration. Because of longevity of Ordovician gastropod genera, the dnorm values are very close to the dtot values. Turnover rates were characterized by two parameters: (a) rate of originations (oi ) per m.y. or rate of extinction (ei ) per m.y., defined as the total number of genera originating or going extinct within the time slice divided by its duration, and (b) mean per capita origination (odi ) or extinction (edi ) rates, defined as the number of originations or extinctions divided by the total number of genera present and by the duration of the time slice in m.y. ■

Classification

Phylogenetic relationships among fossil and living gastropod higher taxa are a widely discussed problem in recent paleontological and neontological literature (Golikov and Starobogatov 1975, 1988; SalviniPlawen 1980, 1990; McLean 1981, 1984; Yochelson 1984; Haszprunar 1988, 1993; Ponder and Warén 1988; Peel 1991a, 1991b; Wagner 1995a; Biggelaar and Haszprunar 1996; Bandel 1997; Nützel 1997, 2002; Ponder and Lindberg 1997; Bandel and Fry´da 1998, 1999; Fry´da and Blodgett 2001, references therein). During the past 20 years great progress in the evaluation of phylogenetic relationships of modern gastropod groups has been made (see Ponder and Lindberg 1997 for review). This progress, based on evaluation of a large number of morphological and anatomic data, has brought relative taxonomic stability to the class Gastropoda. The class may be subdivided (figure 19.1) into Patellogastropoda (= Docoglossa), Archaeogastropoda sensu stricto, Neritimorpha, Caenogastropoda, and Heterobranchia, which are diagnosed by many synapomorphies (see Ponder and Lindberg 1997 for references). The validity of some of these groups is also strongly supported by recent studies based on molecular data (Colgan et al. 2000). Unfortunately, diagnostic features of these higher taxa involve mainly soft-body characters, which could not be used in the fossil record. Thus, there is a relatively stable taxonomic system for living gastropods, which has very limited application to the more numerous fossil gastropods. Fortunately, recent neontological studies have shown that some features

185

= first whorl openly coiled = protoconch II absent = protoconch II present

Eogastropoda

Recent

Triassic

Permian

Silurian

B

Carboniferous

A

Devonian

                  .     Ordovician

Patellogastropoda

????????????

Neritimorpha B

A

Cyrtoneritimorpha

D C

C

D

E

Orthogastropoda

Euomphalomorpha F E

Archaeogastropoda

F F

Perunelomorpha G

Caenogastropoda H

G

Mimospirina H G

Heterobranchia

Apogastropoda

186

FIGURE 19.1. Stratigraphic ranges of main gastropod groups based on protoconch morphology (black bars). Shaded bars show stratigraphic ranges inferred from teleoconch features. Presumed phylogenetic relationships of the higher taxa of gastropods are shown on the right side (see text). Characteristic protoconchs are drawn on the left side. A, larval shell of modern members of the Neritidae. B, larva shell of Late Triassic Pseudorthonychia Bandel and Fry´da 1999. C, Late Ordovician cyrtoneritimorph larval shell. D, Early Devonian cyrtoneritimorph Vltaviela Fry´da and Manda 1997. E, larval shell of the Carboniferous Orthonychia Hall 1843. F, euomphalomorph protoconch of the Early Carboniferous Serpulospira Cossmann 1916. G, Late Silurian perunelomorph larval shell. H, Early Devonian subulitid larval shell. (Sketches made according to Dzik 1994c; Yoo 1994; Fry´da and Manda 1997; Bandel and Fry´da 1999; Fry´da 1999b, 2001.)

such as the nature of early shell development may be applied even in fossil gastropods. Some of modern higher taxa have been identified by their typical protoconch morphology even in Paleozoic strata (see Fry´da 1999b, 2001; Nützel et al. 2000; and Fry´da and Blodgett 2001 for reviews). However, during Early and Mid Paleozoic time there were also several gastropod groups such as the Bellerophontoidea, Macluritoidea, Euomphaloidea, Peruneloidea, Platyceratoidea, Subulitoidea, Helcionelloida, Cyrtoneritimorpha, and Mimospirina, which are not easily connected with modern gastropod higher taxa. ■

Major Groups

The last complete revision of high-level classification of Paleozoic gastropods (Knight et al. 1960) in many cases followed the older classification of Wenz (1938–1944) and placed the great majority of the Ordovician gastropods in the Archaeogastropoda Thiele 1925. Only two superfamilies, Loxonematoidea Koken 1889 and Subulitoidae Lindström 1884, were placed in the Caenogastropoda Cox 1959. The present state of knowledge regarding the definition,

phylogenetic relationships, and paleobiogeography of the major groups of Ordovician gastropods is discussed in the following sections. Such an approach allows us to make a more stable analysis of their diversity patterns.

Patellogastropoda (Eogastropoda) Ponder and Lindberg (1995) divided the class Gastropoda into two groups, Eogastropoda and Orthogastropoda. Eogastropoda, including Patellogastropoda (= Docoglossa), have been considered to represent the first gastropod offshoot (see Ponder and Lindberg 1995, 1997, and references therein). However, there is no undoubted evidence for Paleozoic Patellogastropoda. Until now, the oldest known limpet belonging without doubt to the Patellogastropoda comes from Triassic strata (Hedegaard et al. 1997). Considerations about the nature of shell in the Paleozoic Patellogastropoda divide neontologists and paleontologists into two groups, who dispute whether the ancestors of the Patellogastropoda had anisostrophically coiled shells or bilaterally symmetrical limpet shells like their post-Paleozoic descen-

Gastropods dants (see Ponder and Lindberg 1997). Limpetlike shell morphology was evolved independently among numerous living and fossil groups of Gastropoda, as well as in the Tryblidiida (= Monoplacophora). Thus, it is almost impossible to determine the higher taxonomic position of limpetlike fossil gastropods only on the basis of their teleoconch characters. Some Paleozoic limpets belong to the extinct group Cyrtoneritimorpha (Neritimorpha), with a typical fishhooklike protoconch (Bandel and Fry´da 1999; Fry´da 1999b), and to the Pragoscutulidae Fry´da 1998 (?Caenogastropoda; Fry´da 2001), but the higher taxonomic position of the remaining majority of Paleozoic limpets is still uncertain. Many may belong to the class Tryblidiida (= Monoplacophora). Even if the Paleozoic ancestors of Patellogastropoda had a coiled shell, they still could not be recognized among Paleozoic gastropods.

Archaeogastropoda The concept of the Archaeogastropoda has been changed many times, and different usages may be found even in the most recent studies (e.g., Golikov and Starobogatov 1975; McLean 1981; Bandel 1982; Salvini-Plawen and Haszprunar 1987; Hickman 1988; Haszprunar 1993; Biggelaar and Haszprunar 1996; and Ponder and Lindberg 1997 and references therein). Among the Paleozoic gastropods, the Archaeogastropoda have been considered to be the most commonly occurring gastropod group (figure 19.2C–H). The oldest presumed Archaeogastropoda based on teleoconch morphology are known from the Late Cambrian (Knight et al. 1960; Tracey et al. 1993). The oldest evidence for their occurrence based on the typical protoconch type comes from the Early Devonian (figure 19.1; Fry´da and Bandel 1997; Fry´da and Manda 1997). As mentioned by several authors (see Tracey et al. 1993 and Bandel and Fry´da 1996 for discussion), family-level classification of the Archaeogastropoda requires complete revision and therefore is not used here. As a dominant group within the Early Paleozoic Gastropoda, the Ordovician Archaeogastropoda had a cosmopolitan distribution. In contrast to modern slit-bearing Archaeogastropoda, their Paleozoic representatives occupied a wide range of depositional environments (Peel 1984).

To facilitate analysis of diversity patterns and turnover rates within Ordovician Archaeogastropoda, its component genera are divided into two groups: (1) slit- or sinus-bearing Archaeogastropoda (or “Selenimorpha”) (figure 19.2C–F) and (2) those without this apertural feature (or “Trochomorpha”) (figure 19.2G, H). Ordovician gastropod genera placed by Knight et al. (1960) in any of 22 families of the superfamily Pleurotomarioidea Swainson 1840 are included in the first group together with genera assigned to the superfamily Murchisonioidea Koken 1896. The latter superfamily, which is very common in the Paleozoic, has variously been classified as belonging to the Archaeogastropoda (Knight et al. 1960) or Caenogastropoda (Ponder and Warén 1988) or as a group with uncertain higher position (Tracey et al. 1993). Yochelson (1984) suggested placing the Murchisonioidea close to the Pleurotomarioidea. Recent discoveries of well-preserved protoconchs in Early and Mid Devonian murchisonioideans, including the type genus Murchisonia, provided undoubted evidence that these gastropods belong to the Archaeogastropoda (Fry´da and Manda 1997). The second analyzed group (Archaeogastropoda without apertural sinus or slit) unites Ordovician genera placed by Knight et al. (1960) in Trochonematoidea Zittel 1895, Euomphalopteridae Koken 1896, Holopeidae Wenz 1938, and Microdomatoidea Wenz 1938. Placement of the last two groups within the Archaeogastropoda was supported by protoconch morphology in some of their Mid and Late Paleozoic members (Fry´da and Bandel 1997; Bandel and Nützel pers. comm.). Slit-Bearing Archaeogastropoda (“Selenimorpha”) The oldest slit- or sinus-bearing Archaeogastropoda are known from the latest Cambrian (Tracey et al. 1993). Their generic diversity (figure 19.3B) increased moderately from the beginning of the Ordovician until the Mid to Late Ordovician (TS.4c– 5a), when the generic diversity increased at a much greater rate. The normalized diversity (dnorm ) reaches its highest values (figure 19.3B) during the late Late Ordovician (TS.6a; dnorm close to 20). The number of genera per m.y. (di ) increased during the Ordovician and reached its maximum value (about 12) at the end of the Ordovician. Rate of originations (oi )

187

188

                  .    

FIGURE 19.2. Examples of main groups of Ordovician gastropods and bellerophontiform mollusks. A, B, Tropidodiscus, side views, ×2, late Early Ordovician, Missouri—bellerophontiform mollusks (Amphigastropoda). C, D, Trochonemella, ×1, views of sinus and selenizone, Late Ordovician, Alaska, USNM 422342 and 422347; E, Ectomaria, ×2.6, Late Ordovician, Alaska, USNM 422360; F, Ectomaria, ×2, Late Ordovician, Alaska, USNM422363. C, D, E, and F are examples of the Archaeogastropoda (“Selenimorpha”). G, H, Slehoferia, top and apertural views, ×4, Caradoc, Czech Republic, YA 2602—Archaeogastropoda (“Trochomorpha”). I, Angulospira, ×2, Whiterockian, Nevada, USNM 485257—Mimospirina. J, K, Lytospira, upper and lower surfaces, ×1.3, Whiterockian, USNM 473944; L–N, Malayaspira, ×1.3, Arenig, Malaysia, USNM topotype 473953. J, K, and L–N are examples of the Euomphaloidea (Euomphalomorpha). O, P, Monitorella, apertural and top views ×1, and Q, interior of operculum, ×1.3, Whiterockian, Newfoundland, NFM 321-322—Macluritoidea. R, S, Cyclonema, apertural and side views, ×0.7, Late Ordovician, Tennessee—Neritimorpha (Platyceratidae). T, Subulites (Fusispira), apertural view, ×1.6, Late Ordovician, Alaska, USNM 422371—Caenogastropoda (Subulitoidea). USNM = U.S. National Museum, Washington, D.C.; NFM = Newfoundland Museum, Canada; YA = Czech Geological Survey, Prague.

reached its maximum values close to the TremadocianArenig boundary (TS.2a), early Mid Ordovician (TS.3b), and during the mid Caradoc (TS.5b–c). Rate of extinctions (ei ) was relatively high between the mid Early Ordovician (TS.1d) and early Mid Ordovician (TS.3b), as well as during the Ashgill (TS.6a–c). The Late Ordovician may be characterized as a period of fast radiation of the slit-bearing Archaeogastropoda. The number of genera doubled from the early Late Ordovician (TS.5b) to the late Late Ordovician (TS.6a). This observation fits well

with the fact that the slit-bearing Archaeogastropoda are a dominant gastropod group in Silurian strata. Archaeogastropoda Without Slit (“Trochomorpha”) In contrast to the “Selenimorpha,” there is no reliable evidence for trochomorph Archaeogastropoda in Lower Ordovician strata. During the Mid Ordovician the “Trochomorpha” had low diversity until the beginning of the Late Ordovician, when the number of genera started to increase rapidly, reaching their high-

Gastropods Mimospirina

SUBULITOIDEA

A

NERITIMORPHA LOXONEMATOIDEA ARCHAEOGASTROPODA-Trochomorpha PERUNELOIDEA CYRTONERITIMORPHA

D

dtot

dtot 2 1o dnorm ei di 0 0 i

dnorm di

MACLURITOIDEA MIMOSPIRINA EUOMPHALOMORPHA ARCHAEOGASTROPODA - Selenimorpha

1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c

Archaeogastropoda

dtot

Selenimorpha

dtot 5 1 dnorm oei di 0 0 i

1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c

Euomphalomorpha

B

E

dtot

dtot 3 1 o dnorm ei di 0 0 i

dnorm dnorm

di

FIGURE 19.3. A, stratigraphic ranges of the main Ordovician gastropod groups. B–F, generic biodiversity and origination and extinction rates of Archaeogastropoda, Mimospirina, Euomphalomorpha, and Macluritoidea. Lines show the diversity (dtot = total diversity, dnorm = normalized diversity; and di = number of genera per m.y.). Bars show origination and extinction rates (white = rate of originations [oi ], black = rate of extinctions [ei ]). For definition of these parameters, see text.

di 1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c

Archaeogastropoda

dtot

Trochomorpha

dtot 3 dnorm di 0

1 0

oi ei

dnorm

1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c LOWER ORDOVICIAN

DARRIWILIAN

489 Ma

M ORDOVICIAN

472

460.5

di

1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c DARRIWILIAN

TREMADOCIAN

UPPER ORDOVICIAN

F

dtot

dtot 2 1 o dnorm ei di 0 0 i

dnorm di

TREMADOCIAN

1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c

Macluritoidea

C

M ORDOVICIAN

LOWER ORDOVICIAN

443 489 Ma

472

est value in the late Late Ordovician (figure 19.3C). Arenig (TS.3a–4a), early Late Ordovician (TS.5a–b), and early Ashgill (TS.6a) are time intervals with the highest origination rate. The pattern of diversification in Mid and Late Ordovician “Trochomorpha” fits well with that of the “Selenimorpha.”

Mimospirina Mimospirina was established by Dzik (1983) as a new suborder of the Archaeogastropoda uniting members of the extinct families Clisospiridae and Onychochilidae (figure 19.2I). Opinions on the higher taxonomic position of both families have often changed (see Fry´da 1992 and Fry´da and Rohr 1999 for review). McLean (1981) suggested that the members of the superfamilies Macluritoidea and Clisospiroidea do not belong to his suborder Euomphalina but represent lineages apart from this group. On the other hand, Linsley and Kier (1984), on the basis of a functional analysis, proposed uniting the Clisospiroidea, Macluritoidea, and possibly the Euomphaloidea into their new class Paragastropoda. The class Paragastropoda was considered to represent untorted mollusks.

UPPER ORDOVICIAN

460.5

443

Discoveries during the past decade revealed new data on the early shell morphology of several gastropod groups, which have enabled a reevaluation of the phylogenetic relationships of the Mimospirina. Quite differing protoconch morphology of members of the Clisospiroidea (Dzik 1983; Fry´da and Rohr 1999) and Euomphaloidea (Yoo 1994; Bandel and Fry´da 1998; Nützel 2002) suggests that the class Paragastropoda is an artificial group with no zoological validity (Fry´da 1999b). These facts also support Dzik’s (1983) concept of the Mimospirina. The multiwhorled protoconch in Mimospira represents a true larval shell (protoconch II), and its large size and shape suggest the presence of a nonplanktotrophic larval stage during which the larval shell was formed. This interpretation argues against the inclusion of Mimospirina within the Archaeogastropoda (Fry´da 1999b, 2001). The Mimospirina are relatively rare in CambrianDevonian marine communities. The highest diversity of the Ordovician clisospirids and onychochilids is found in the Baltic fauna. Paleogeographic distribution of the Ordovician Mimospirina was summarized by Fry´da and Rohr (1999).

189

190

                  .     The generic diversity of the Mimospirina was relatively low up to the early Mid Ordovician (TS.3b). From approximately the beginning of the Darriwilian (TS.4a), the normalized diversity (dnorm ) started to increase strongly, attaining its highest value in the middle of the Late Ordovician (figure 19.3D). The number of genera per m.y. (di ) increased during the Ordovician and reached its maximum value in the later Late Ordovician (TS.6b). The latest Ordovician (TS.6c) represents a time of sharp decrease in both normalized diversity and number of genera per m.y. The rate of originations (oi ) is highest at the beginning of the Darriwilian (TS.4a) and in the late Late Ordovician (TS.6a), followed by a distinct extinction event in the latest Ordovician (TS.6b–c).

Euomphaloidea (Euomphalomorpha) Yochelson (1956) interpreted the Euomphaloidea (figure 19.2J–M) as derived from the Macluritoidea in Early Ordovician time and assigned three families (Euomphalidae Helicotomidae, and Omphalotrochidae) to this superfamily. This concept was followed by Knight et al. (1960) and many paleontologists subsequently. McLean (1981) considered the deepsea hot vent limpet Neomphalus McLean 1981 to be related to extinct Euomphaloidea. For this reason, McLean (1981) united members of the modern Neomphaloidea with the Paleozoic Euomphaloidea and placed them in the Archaeogastropoda. Yoo (1994) and Bandel and Fry´da (1998) found an unusual protoconch morphology in some Devonian and Carboniferous euomphaloidean genera, which distinguishes them from members of extant gastropod groups (Patellogastropoda, Archaeogastropoda, Neritimorpha, Caenogastropoda, and Heterobranchia). For this reason, Bandel and Fry´da (1998) established a new taxon, Euomphalomorpha, which is considered to represent an independent gastropod group, known only from the Paleozoic (Cambrian–Permian). Recently, Nützel (2002) confirmed earlier observations (Yoo 1994; Bandel and Fry´da 1998) on the nature of the protoconch and shape of the boundary between protoconch and teleoconch. He also noted some affinities among Euomphaloidea, Docoglossa, Cocculiniformia, and Neomphalidae. In the Ordovician Euomphalomorpha we include the following families:

Euomphalidae de Koninck 1881, Helicotomidae Wenz 1938, and Ophiletidae Knight 1956. The Euomphalomorpha are one of the dominant groups of Paleozoic gastropods. During Ordovician time they are known mainly from paleotropical regions. Their highest biodiversity is found in carbonatedominated facies of Laurentia and Baltica. Euomphaloidea (figures 19.1, 19.3E) are known from the Late Cambrian to the Late Permian. During the Early and early Mid Ordovician the number of genera per m.y. (di ) slowly increased and reached its maximum value in the early Mid Ordovician (TS.3b and 4a). Subsequently, the number of general per m.y. dropped (TS.4b) and remained approximately constant until the end of the Ordovician (figure 19.3). The curve of the normalized diversity (dnorm ) shows two peaks (early Arenig, TS.2b, and early Mid Ordovician, TS.3b–4a). Both the rates of originations (oi ) and extinctions (ei ) are generally higher during the Early and early Mid Ordovician than in the later Ordovician (figure 19.3E). The beginning of Arenig (TS.2a) and its mid to later part (TS.3a–b) are times with the highest origination rates. On the other hand, the rate of extinctions (ei ) is highest during the early Darriwilian (TS.4a) and latest Ordovician.

Macluritoidea Although Macluritoidea (figure 19.2N, O) belong to the best-known group of Ordovician gastropods (Rohr 1979, 1994; Rohr and Gubanov 1997), little is known about their higher taxonomic position. Knight et al. (1960) placed Macluritoidea Fisher 1885, together with Euomphaloidea de Koninck 1881, in the suborder Macluritina Cox and Knight 1960 of the Archaeogastropoda. The concept of Macluritina in uniting Macluritoidea and Euomphaloidea was later criticized by several authors (Morris and Cleevely 1981; Dzik 1983; Linsley and Kier 1984; Yochelson 1984). As shown by numerous studies (Rohr 1979; Blodgett et al. 1987; Rohr et al. 1992; Gubanov and Rohr 1995; Gubanov and Tait 1998), Macluritoidea lived in warm, shallow marine waters in both carbonate and siliciclastic facies and are known from all Ordovician paleocontinents situated in tropical regions. More detailed data on the paleogeographic distribu-

Gastropods tion of the Ordovician Macluritoidea may be found in Gubanov and Rohr (1995). The oldest Macluritoidea (in the sense of Gubanov and Rohr 1995) are known from the early Arenig (TS.2a). They are restricted to the Ordovician and reached their highest diversity in the early Mid Ordovician (TS.3b–4a). In the mid Darriwilian (figure 21.3F) their generic diversity rapidly dropped and remained low until the Ashgill, when the Macluritoidea became extinct. The pattern of diversification of the Macluritoidea resembles that of the Ordovician Euomphaloidea.

Neritimorpha Bandel (1982) pointed out that the Neritimorpha Golikov and Starobogatov 1975 represent an independent gastropod group characterized by unique protoconch morphology. This opinion seems to be supported by the study of gastropod cleavage patterns (Biggelaar and Haszprunar 1996) and by cladistic analysis of a large number of morphological and anatomic data of modern gastropods (Ponder and Lindberg 1997). According to the latter authors, the Neritimorpha represent a sister group to Archaeogastropoda and Apogastropoda (Caenogastropoda and Heterobranchia). Besides several synapomorphies including mainly soft body characters (see Ponder and Lindberg 1997), the modern and fossil Neritimorpha may be characterized by a typical, strongly convolute protoconch (Bandel 1982). The oldest undoubted evidence for this protoconch type is known from Triassic strata (Bandel and Fry´da 1999). Presumed Paleozoic members of Neritimorpha (see Knight et al. 1960) have two different types of protoconch (figure 19.1): (1) closely coiled (but not convolute) protoconch (e.g., Platyceratidae, Plagiothyridae, Naticopsidae, and Oriostomatoidea), and (2) openly coiled, fishhooklike protoconch (Vltaviellidae and Orthonychiidae). The first group may represent ancestors of modern Neritimorpha (Cycloneritimorpha). Presumed Paleozoic members of Neritimorpha developing the fishhooklike larval shell (figure 19.1) were united into the taxon Cyrtoneritimorpha (Fry´da 1998, 1999b; Bandel and Fry´da 1999; Fry´da and Heidelberger 2003). The Ordovician Neritimorpha (figure 19.2P, Q) are known from tropical as well as high-latitude re-

gions. Ordovician Cyrtoneritimorpha are currently known only from Baltica (Bockelie and Yochelson 1979; Dzik 1994c). On the basis of their teleoconch shape, the first presumed Neritimorpha (Platyceratidae and Oriostomatidae) appeared during the Mid Ordovician. On the other hand, isolated protoconchs of the Cyrtoneritimorpha are known in upper Lower Ordovician strata (Bockelie and Yochelson 1979). Dzik (1994c: figure 22a) illustrated cyrtoneritimorph protoconchs also in Middle Ordovician strata. Thus, the existence of Cyrtoneritimorpha in the Ordovician is beyond doubt, but there is no certain record of Neritimorpha with a strongly convolute protoconch (see discussion in Bandel and Fry´da 1999). Because of the lack of data on Ordovician neritimorph gastropods, a detailed quantitative analysis of their diversity pattern and turnover rates would be meaningless.

Caenogastropoda (Subulitoidea and Peruneloidea) Salvini-Plawen and Haszprunar (1987) introduced the term Apogastropoda for the Caenogastropoda and basal Heterobranchia. Haszprunar (1988) suggested that the first caenogastropods and heterobranchs had a common ancestor. Although there is a rich fossil record for the Late Paleozoic Caenogastropoda (see Nützel et al. 2000 for references), evidence for pre-Carboniferous Caenogastropoda and Heterobranchia is very poor (Fry´da and Blodgett 2001). The discovery of the first presumed preCarboniferous caenogastropod was reported from the Early Devonian (Pragoscutula wareni Fry´da 1998; see Fry´da 2001: figure 1). The true larval shell (protoconch II) of the Caenogastropoda was also recently documented in some members of the Devonian (Fry´da 2001: figures 3, 4) and Carboniferous (Nützel et al. 2000) Subulitoidea. The latter superfamily includes true Caenogastropoda with characteristically closely coiled larval shells. On the other hand, some Devonian gastropod genera closely resembling subulitoidean gastropods (Chuchlinidae of Peruneloidea Fry´da and Bandel 1997) developed true larval shells with an openly coiled first whorl (Fry´da and Manda 1997; Fry´da 1999a). According to Fry´da (1999b, 2001), the Ordovician-Carboniferous Peruneloidea

191

192

                  .     may be considered to be an independent gastropod group at the ordinal level (Perunelomorpha), which represents the ancestral group (or basal group) of the Caenogastropoda or even the whole Apogastropoda (figure 19.1). Ordovician Subulitoidea (figure 19.2R) are known mainly from tropical regions (Laurentia and Baltica). Distribution of the Ordovician Peruneloidea is imperfectly known, with the only available data from Baltica (Bockelie and Yochelson 1979; Hynda 1986; Dzik 1994c). The presumed members of the Subulitoidea Lindström 1884 developed two types of protoconchs: (1) closely coiled larval shells like modern Caenogastropoda (Nützel et al. 2000; Fry´da 2001), or (2) larval shells with an openly coiled first whorl (Peruneloidea; see Fry´da 2001 for references). The oldest members of the first group are known from the Early Devonian. However, members of the second group are known to be present from the Early Ordovician to the Carboniferous (figures 19.1, 19.3A). Unfortunately, no information exists concerning the morphology of the protoconch of the type genus Subulites Emmons 1842, and so the relationships of the Subulitoidea and Peruneloidea cannot be resolved. Regardless of this taxonomic problem, both of the latter groups are well known from the Ordovician (the first on the basis of teleoconchs and the second by the discovery of typical protoconchs). Both gastropod groups developed true larval shells, and their morphology supports their assignment to the Caenogastropoda. Like the Neritimorpha, there is good qualitative evidence for the existence of the Caenogastropoda (or their ancestors) in the Ordovician (figures 19.3A, 19.4D), but the lack of data precludes a detailed quantitative analysis of their diversity pattern. ■

Discussion

To date, only a few papers (Sepkoski 1995; Wagner 1995b) have investigated diversity patterns among the Ordovician gastropods. As in these studies, we found that their diversity (figure 19.4A) was increasing from the earliest Ordovician until the Late Ordovician, when it dropped drastically close to the Ordovician/Silurian boundary. In contrast to previous studies, our analysis has been based on much

GASTROPODA Generic diversity

dtot dnorm

20

0

di A

1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c

Rates of originations (oi) and extinctions (ei) 4

ei

oi 0

B

Rates of originations (odi) and extinctions (edi) 0.2

0

C

odi

Amphigastropoda

edi Mimospirina Caenogastropoda Neritimorpha

Archaeogastropoda Trochomorpha

Euomphaloidea

Selenimorpha

D

Macluritoidea 1a 1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c 5a 5b 5c 5d 6a 6b 6c

FIGURE 19.4. A–C, generic diversity (dtot , dnorm , and di ) of the Ordovician gastropods and their rates of originations (oi and odi ) and extinctions (ei and edi ). (For definition of these parameters, see text.). D, relative richness of the major gastropod taxa.

more detailed stratigraphic data, which has revealed some new and more detailed observations. The mid–Late Ordovician (TS.5c–6b) is the time of highest diversity. Another peak of high diversity is found in the lowermost Darriwilian (TS.4a). Both the rates of originations (oi and odi ) show four distinct, regularly spaced peaks (figure 19.4B, C). More detailed analysis of this phenomenon shows that each of these peaks has contributions from different higher taxa coming from different regions. A relatively high rate of origination (see odi in figure 19.4C)

Gastropods also occurs at the base of the Tremadocian. Unfortunately, both very low total diversity and the lack of exact stratigraphic data for Tremadocian gastropod distribution preclude a more detailed analysis. The extinction rates through the Ordovician show a much simpler pattern. The highest value was reached close to the Ordovician/Silurian boundary. This result is in full concordance with previous observations (Sepkoski 1995; Wagner 1995b). Our analysis has also revealed a very strong peak of the extinction rates at the beginning of the Darriwilian (TS.4a). The bioevents mentioned earlier are discussed in detail in the next sections.

Early Arenig Origination Event The first major peak of origination rates recognized in the Ordovician coincides with the base of the Arenig (TS.2a—graptolite-based Tetragraptus approximatus Zone). Both rates of originations (oi and odi ) reached maximum values, 8.5 and 0.5, respectively. The increase in generic diversity as well as changes in generic composition of both the slitbearing Archaeogastropoda and Euomphaloidea made the most significant contribution to this peak (figure 19.4B, C). This faunal change is coupled with the latest Tremadocian (TS.1d) extinction event, during which 11 genera became extinct. Barnes et al. (1995) noted a first-order regression/transgression couplet close to the Tremadocian-Arenig boundary. A major regression could be responsible for the extinction event, which was followed by a period (early Arenig, TS.2a) of adaptive radiation during rapid transgression. Barnes et al. (1996) suggested that during transgression the shallow platform waters attained a higher level of oxygenation, which allowed the rapid radiation of a new Ordovician fauna, and this fauna became differentiated from the Cambrian fauna, which was adapted to relatively low levels of oxygenation. Currently, it is impossible to test this hypothesis because of a lack of data on the environmental distribution of Late Cambrian and Early Ordovician gastropods. Nevertheless, besides the trilobites, graptolites, and conodonts (Barnes et al. 1996), gastropods are an additional animal group showing significant faunal change at the base of the Arenig (TS.2a— T. approximatus Zone).

Earliest Darriwilian Origination/Extinction Events The interval close to the lower boundary of the Darriwilian exhibits high turnover rates (figure 19.4). Origination rates (oi and odi ) reached maximum values just before the Darriwilian (TS.3b) and remained high during the earliest Darriwilian (TS.4a). At the same time, the extinction rates increased strongly to a local maximum (early Darriwilian, TS.4a). This couplet of earliest Darriwilian origination/extinction events corresponds to the most distinct faunal change within the Ordovician (28 new gastropod genera originated, and 14 genera became extinct). The earliest Darriwilian origination event influenced the diversity of all Ordovician gastropod groups (figures 19.3, 19.4). The first representatives of more advanced gastropod groups (Loxonematoidea, Subulitoidea, figure 19.3A) also appeared during this origination event. The earliest Darriwilian origination/ extinction events probably had a crucial influence on the subsequent evolution of the Euomphaloidea and Macluritoidea. The latter superfamilies reached their highest diversity close to the lower boundary of the Darriwilian (TS.3b–4a). Subsequently their diversities decreased strongly (i.e., 6 of 11 euomphaloidean genera became extinct). During the rest of the Ordovician the proportion of both Euomphaloidea and Macluritoidea to the total gastropod diversity decreased (figure 19.4). The former superfamily underwent the next of its several radiation events during the Mid and Late Paleozoic. On the other hand, the Macluritoidea definitely became extinct during the mass extinction event close to the Ordovician/ Silurian boundary. The similarity in the diversity and turnover patterns of the Macluritoidea and Euomphaloidea (figure 19.3E, F) may suggest similarities in some of their life strategies or perhaps some of their phylogenetic links (see discussion earlier in this chapter). A similar though slightly younger bioevent (the so-called Basal Llanvirn Bio-Event) was described by Barnes et al. (1996), which distinctly influenced the diversity of trilobites, graptolites, and conodonts. On the other hand, Sepkoski (1995) pointed out that the Ordovician diversification of typical Paleozoic groups really started in the late Arenig (i.e., in the earliest

193

194

                  .     Darriwilian). The data here provide evidence that the Gastropoda also belong to these groups. Barnes et al. (1996) noted that at the level of the so-called Basal Llanvirn Bio-Event a major excursion in the 87Sr/ 86Sr isotope curve occurred. These authors suggested a possible relation between the bioevent and the increased seafloor spreading and ridge activity inferred from the 87Sr/86Sr isotope curve. However, new data on marine 87Sr/86Sr isotope evolution (e.g., Veizer et al. 1999) show that the decrease in the 87Sr/86Sr values started much earlier (latest Tremadocian) and continued up to the Late Ordovician. In contrast to that, our analysis reveals a short period of very strong increase in extinction rates at the beginning of the Darriwilian (TS.4a). An increase of extinction rates at this time was previously observed in some other faunal groups (e.g., Sepkoski 1995). It is difficult at this time to speculate on the reasons for the early Darriwilian extinction event. However, it is interesting to note that the age of the relatively large Ames impact structure (about 470 Ma) in Laurentia (Koeberl et al. 2001) agrees well with the age of the extinction event. Future detailed paleontological, climatic, and isotopic studies may shed new light on the reasons of this bioevent.

Ashgill Origination/Extinction Events After a short stasis in the middle of the Late Ordovician (figure 19.4), the origination rates increased again from the beginning of the Ashgill (TS.6a). During this origination event, 17 new genera appeared, belonging mainly to the Archaeogastropoda and Mimospirina. After this event, no new generic taxa have been documented as originating in the remainder of the Ordovician. On the other hand, extinction rates (ei and edi ) started to increase dramatically until they peaked at the end of the Ordovician (figure 19.4B, C). During the Ashgill 38 gastropod genera disappeared, and this crisis affected all gastropod groups. A detailed analysis of extinctions in different environments and paleogeographic regions should be the subject of a future study. At present it should be noted that the exact time of the highest extinction rate for each gastropod group seems to differ slightly (figure 19.3), even though the extinction curve for all gastropods (figure 19.4B, C) shows a simple ascending pattern through the Ashgill. This may reflect multiple crises during the uppermost Ordovician as suggested by Copper (2001c). ■

Conclusions

Early Caradoc Origination Event The early Caradoc (TS.5a–b) also represented an interval with a distinct rise in origination rates (figures 19.3, 19.4). In contrast to the previous bioevents, it is not coupled with a significant extinction event. The typical feature of this bioevent is the considerable acceleration of the radiation (31 gastropod genera had their first appearances, and only 5 genera became extinct) among the majority of gastropod groups (“selenimorph” and “trochomorph” Archaeogastropoda, Mimospirina, Neritimorpha, Amphigastropoda, and Caenogastropoda) after the late Darriwilian stasis. Several authors (see Barnes et al. 1995; Sepkoski 1995; and Bassett et al. 1999c for reviews) have earlier mentioned an origination bioevent that occurred during the early Caradoc. Barnes et al. (1996) suggested that the latter bioevent corresponds to a significant regressive phase followed by mid Caradoc transgression inundating most cratons.

During the past several years views on the dynamics of evolution have changed markedly (see Brett and Baird 1995 and Sheehan 2001a for references). It seems now that there were geologically long intervals of faunal stability (stasis), which lasted several m.y. and display little taxonomic and ecologic change. These periods of stability were bounded by shorter periods of evolutionary and ecologic turnover, characterized by high levels of speciation and extinction. These evolutionary patterns have been documented for many marine animal groups (e.g., Boucot 1975, 1983; Patzkowsky and Holland 1997). Although our analysis of diversity and turnover of the Ordovician gastropods also revealed a similar pattern (figure 19.4), it shows that the periods of stability were of lengths similar to the intervals with high turnover rates, in contrast to the studies of many other marine groups. This pattern may be because gastropods belong to the groups with lower turnover rates (Sepkoski 1995). It could also be that this pattern is an artifact of the

Gastropods poor quality of stratigraphic data available for some Ordovician gastropod taxa. Nevertheless, disregarding relative lengths of stases, it is beyond doubt that the evolutionary pattern of the Ordovician gastropods shows periods of relative stability separated by periods with high levels of turnover rates. The Ordovician was a time of substantial growth in gastropod taxonomic richness. Even though our knowledge of the Tremadocian gastropod fauna is still poor, this time interval may be characterized by relatively high origination rates. The first significant extinction event was observed at the end of Tremadocian (TS.1c). The following early Arenig origination bioevent (TS.2a—Tetragraptus approximatus Zone) represents the most important Ordovician origination affecting the gastropods. The subsequent earliest Darriwilian origination/extinction couplet corresponds to the most distinct generic change within the Ordovician, causing significant acceleration of development and radiation in some gastropod groups (Archaeogastropoda, Mimospirina, Neritimorpha, Amphigastropoda, and Caenogastropoda?) as well as a crisis for the Euomphaloidea and Macluritoidea. The earliest Darriwilian origination/extinction bioevents also considerably changed the relative abundance of the higher gastropod taxa (figure 19.4D). After the mid to late Darriwilian stasis, the early Caradoc origination bioevent (TS.5a–b) again dis-

tinctly accelerated radiation of the majority of gastropod groups. In contrast to the previous bioevent, no significant changes in the relative abundance of higher gastropod taxa were observed during the early Caradoc and the early Ashgill origination bioevents. The latest Ordovician was a time of the highest extinction rates, affecting all the gastropod groups. For a better understanding of the biodiversity patterns of Ordovician gastropods, future efforts should be focused on the study of their species-level diversity, as well as on a comparison of their diversities and turnover rates in different environments and paleogeographic regions. This however, cannot be attained without future intensive, specimen-based studies.

 This study was supported by the Grant Agency of the Czech Republic (grant 205/01/0143), Alexander von Humboldt-Stiftung, and IGCP project no. 410. Rohr’s work was supported in part by grants from the National Geographic Society, Committee for Research and Exploration, and from Sul Ross State University, Faculty Research Enhancement Program. We also acknowledge Ian Percival (Sydney), Arthur J. Boucot, and Robert B. Blodgett (Oregon) for their helpful, critical reviews of this chapter.

195

20

Bivalve and Rostroconch Mollusks John C. W. Cope

B

ivalves and rostroconchs are often similar in shape, and some rostroconchs were identified originally as bivalves. The apparently bivalved condition of the rostroconchs is, however, illusory, as they develop from a univalved protoconch as opposed to the bivalved protoconch of bivalves. Rostroconchs are therefore pseudobivalved; they appear to have developed from laterally compressed monoplacaphorans in the Early Cambrian and are believed in turn to have given rise to the bivalves, also in the Early Cambrian. The two groups share a laterally compressed body organization that clearly had a relatively straight gut with mouth and anus well separated at the anterior and posterior of the shell, respectively. The molluscan head was very poorly developed or absent, and sensory functions were concentrated in the mantle margins. Although there is no radula in bivalves, it is possible that some rostroconchs had this rasping tonguelike organ. The similarities between these two classes led Runnegar and Pojeta (1974) to combine them in the subclass Diasoma, together with the class Scaphopoda. The latter class has been recorded in the past from the Ordovician, but more recent work has disproved these Ordovician records, and it is now clear that the scaphopods were derived from the conocardioid rostroconchs in the Devonian (Engeser and Riedel 1996).

196

■

Bivalvia

Bivalves are one of the least well-known groups of Ordovician fossils, primarily because they are often extremely rare, especially in the Early and Mid Ordovician; however, in particular facies and at individual times they become locally the dominant invertebrate group, even from the Early Ordovician (Cope 1996b). Although a handful of taxa are known from Lower and Middle Cambrian rocks, there are no undoubted Upper Cambrian bivalves, and thus all Lower Ordovician forms are cryptogenic. Pojeta (1980) figured a possible Upper Cambrian bivalve from Maryland, but this specimen shows no features unique to bivalves, and Hinz-Schallreuter (2000) has pointed out that Berg-Madsen’s (1987) claimed Upper Cambrian specimen of Tuarangia could equally be from the Middle Cambrian, from which the genus was already known. Pojeta (1971) recorded some 1,400 species of Ordovician bivalves, but in compiling the data for this chapter it has not proved possible to use species, since many of the primary descriptions (particularly of Laurentian species) are now in need of major revision. The specific names include many synonyms, and many species were inadequately founded. Some genera, too, are ill founded, and without reexamination of the type material, doubts exist concerning their interpretation.

Bivalve and Rostroconch Mollusks

Pteriomorphia

Heteroconchia

Anomalodesmata

AUTOLAMELLIBRANCHIATA

Trigonioida

Nuculoida

PROTOBRANCHIA

Solemyoida

This problem has been compounded by assigning bivalves from other paleocontinents to these doubtful taxa. The same criticism can be leveled at many of the forms figured in the nineteenth century from classical areas of Europe including Bohemia and Brittany, and only since the latter part of the twentieth century have monographic descriptions been more rigorous and included details of the number of specimens, their repository and registered number, and the precise age. From such data useful analysis can result, and it can now be shown that latitude-dependent differences exist in Ordovician bivalve faunas (Babin 1993, 2000; Cope and Babin 1999; Cope 2002). Some generic names have been so widely applied that they have become meaningless. A prime example is the nuculoid Ctenodonta. The Upper Ordovician type species, C. nasuta (Hall), is very distinctive, but the name has been applied to nuculoids of totally different shape and to those lacking musculature or dentition; Pojeta (1971) recorded that 183 species had been assigned to the genus. Herein all records of the genus prior to those from the Middle Ordovician, and those that show no kinship to the type species, have been discounted. Genera that remain uninterpretable through inadequate original description, such as Allodesma Ulrich 1894, are also omitted. Specific names, too, cause problems because of the differing levels of treatment by various authors. The description of 20 species belonging to one genus from one locality would not be accepted today, but this is the level of splitting adopted by more than one author in the nineteenth century and first half of the twentieth. Pojeta (1971) noted that of the 1,400 specific names he had recorded for Ordovician bivalves, more than 700 were accounted for by only 16 genera; thus true estimates of the number of species are hard to make, and a “best guess” approach has been used to try to determine the specific diversity as it would be interpreted today. Collection failure remains a significant factor in bivalve diversity calculations and is evident in several groups. For instance, the earliest arcoidean, Catamarcaia, is known from the “Middle Arenig” of Argentina (TS.3a); the next youngest arcoidean is from the Lower Llandovery, demonstrating a total lacuna in early arcoidean history during the Mid and Late Ordovician.

Cardiolarioidea

FIGURE 20.1. Classification of the Bivalvia adopted herein, showing the probable phylogenetic links between the major bivalve groups. The two major subdivisions (horizontal) are subclasses, and the vertical groups are superorders. The superfamily Cardiolarioidea is at present not assigned to any higher-level taxon. After Cope (2002). Reproduced by permission of the Geological Society of London.

The classification of bivalves used herein is that of Cope (2000, 2002), depicted in figure 20.1. Lower Ordovician faunas are very rare, Tremadocian faunas (TS.1a–d) being certainly known only from Argentina (Harrington 1938), the Montagne Noire (Babin 1982), Central Australia (Pojeta and Gilbert-Tomlinson 1977), and Morocco (pers. observ. 2001). By the early Arenig (TS.2a) bivalve faunas are known from six localities worldwide; this early Arenig increase in abundance was mirrored by increases in diversity and size. Cope (1995) correlated this with the evolution of the feeding gill and identified a group of bivalves, the Cardiolarioidea, that links the protobranch bivalves with the autolamellibranch forms in which the gills are used for feeding as well as respiration (Cope 1995, 1997b). Early Ordovician bivalves are confined entirely to the Gondwanan continent, including Avalonia (Cope and Babin 1999), although they range from very high latitudes (e.g., Morocco) to equatorial latitudes (e.g., Central Australia). There are 37 recorded genera and an estimated 60 Early Ordovician species. All the genera are new,

197

198

 . .  and the explosive nature of the Early Ordovician bivalve diversification can be gauged from the fact that these genera belong to no fewer than 18 families. The major diversification period was in the early Arenig (TS.2a), the most important faunas coming from South Wales, with 20 species belonging to 18 genera (Cope 1996b), and the Montagne Noire, with 9 species belonging to 9 genera (Babin 1982); this represents a major advance over the total of probably 6 known species worldwide in the late Tremadocian (TS.1d). By the Mid Ordovician bivalves had become much more abundant locally; Babin and Gutiérrez-Marco (1991) recorded them from 87 Spanish localities, but elsewhere around the Gondwanan margins, as in Avalonia, bivalves remained generally rare, and Cope (1999) described Anglo-Welsh Middle Ordovician bivalves essentially from two localities. Worldwide, Middle Ordovician bivalves are known from more than 100 localities. Bivalves were still largely restricted to Gondwana and Avalonia in the Middle Ordovician (Cope and Babin 1999); unequivocal Middle Ordovician bivalves are otherwise known only from Baltica (two species) and the margin of Laurentia on Svalbard (one species). The many bivalve records from the Middle Ordovician of the remainder of Laurentia, Siberia, and Bohemia have all proved to belong to the Upper Ordovician (TS.5a or later). Kazakhstan, however, has three latest TS.4c species (L. E. Popov pers. comm.). In the Middle Ordovician the number of genera increased to 53, of which 35 are new and 18 are holdover genera; these include an estimated 120 species. The 18 recorded families include 4 new ones, but 3 families known from both the Lower and the Upper Ordovician are as yet unknown from the Middle Ordovician. In the Late Ordovician, bivalves finally reached the Laurentian and Baltoscandian carbonate platforms, providing ideal epifaunal habitats where a second major radiation took place (Cope and Babin 1999). Greatest diversification occurred in pteriomorph groups, but the exclusively infaunal nuculoids also diversified more here, perhaps taking advantage of the warmer low-latitude waters (Cope 2002). There was a dramatic increase to 93 genera (63 new) that include an estimated 350–500 species, though exact numbers are hard to establish. The 20 families include

only 2 new ones; thus familial diversity remained little different throughout the Ordovician Period after the early Arenig (TS.2a) explosive diversification. Forty-nine genera (53 percent) are pteriomorphians, and 24 genera (26 percent) are nuculoids; by far the largest number of species is also among the pteriomorphians and may amount to three-quarters of Upper Ordovician species. Upper Ordovician bivalve faunas are dominated by low-latitude faunas. Few species are known from high latitudes in the Late Ordovician, as the faunas are poorly known. In contrast, Laurentian Upper Ordovician faunas are in need of major taxonomic revision to reinterpret the material and arrive at a realistic number of species.

Nuculoids Nuculoids are represented by nine genera described hitherto from the Lower Ordovician, although some, originally assigned to Ctenodonta, require new generic names. The earliest are present in the basal Ordovician in Argentina (TS.1a—Harrington 1938) and are possibly indirect descendants of the Cambrian taxodont genera Pojetaia and Tuarangia, both of which occur in the Middle Cambrian. Most of the Lower Ordovician species are from high latitudes. In the Middle Ordovician the number of genera so far described is 17; of these, 13 are new and 4 holdover genera. Four genera are known only from low-latitude Gondwana. In the Upper Ordovician there are 24 nuculoid genera, of which 14 are new and 10 holdover (figure 20.2); the new genera are principally lowlatitude forms from Laurentia, Siberia, and Gondwana, reflecting the greater diversity of nuculoids at low latitudes (Cope 2002).

Solemyoids Solemyoids are anteriorly elongate protobranchs that are among the rarest Ordovician bivalves, and there is controversy about their origins. Cope (1996b) described the earliest known form, Ovatoconcha, from the early Arenig (TS.2a) of South Wales, but previously Pojeta (1988) had postulated an origin from ctenodontid nuculoids in the Late Ordovician and assigned the genera Dystactella and Psiloconcha to the group. As pointed out by Cope (2000), however,

indet.

Psiloconcha

Praeleda Lyrodesma Pseudarca Brachilyrodesma

Inaequidens Tromelinodonta Noradonta

Eritropis

Cardiolaria

Trigonoconcha Hemiconcavodonta Emiliania Concavoleda Villicumia Dystactella

Tancrediopsis

Sluha Homilodonta Cuyopsis Tironucula Natasia

Ekaterodonta

FIGURE 20.2. Range chart of the genera of Ordovician Bivalvia belonging to the Nuculoida, Solemyoida, Cardiolarioidea, and Trigonioida. In most cases the actual range within a time slice is not known, and the range is taken to the nearest upper and lower timeslice boundaries. Dotted lines signify ranges for which a genus must have survived but for which no record yet exists. Higher level taxa shown in subheadings are orders (in italics) and superfamilies (in roman).

1c

1a

indet.

1b

praenuculid

Tremadoc

Ibexian

1d

praenuculid

2a

Ovatoconcha

2b

Thoralia Pensarnia Paulinea

Synek

Arenig

ORDOVICIAN

2c

Praenucula

Lophoconcha

4a

3a

TREMADOCIAN

LOWER

4b

3b

472 Ma

489 Ma

4c

Aloconcha

Whiterockian Llanvirn

DARRIWILIAN

MIDDLE ORDOVICIAN

5a 460.5 Ma

Trigonioida

Solemyoida

Suria

5b

Afghanodesma

5c

Fidera Nuculites Palaeoneilo Concavodonta Johnmartinia Similodonta Sthenodonta Palaeoconcha Sibiroctenia Arcodonta Zeehania Myoplusia Cadomia

Ashgill

5d Ctenodonta

UPPER

6b 6a

Mohawkian Caradoc

ORDOVICIAN

Cincinnatian

6c

Nuculoida

Costaledopsis

443 Ma

Cardiolarioidea

TIME SLICES

North American & British series

GLOBAL SERIES & STAGES

Bivalve and Rostroconch Mollusks

the earliest true Ctenodonta species, from the Middle Ordovician, all have umbones that are well to the anterior, and all lack ligamental nymphs that both Pojeta (1988) and Waller (1990, 1998) believed to be synapomorphies linking the ctenodontids and the solemyoids; thus ligamental nymphs originated independently in the two groups, and the claimed intermediate genus Dystactella is more likely to be a persistent intermediate stock, with the true origins of the solemyoids lying unknown in the earliest Ordovician (or latest Cambrian). There are so far no solemyoids known from the Middle Ordovician.

Cardiolarioids The Cardiolarioids are a small group of taxodont bivalves with hinges designed for wider valve opening while maintaining closer articulation of the valves than is possible with nuculoid hinges. Cope (1995) suggested that this hinge design was directly linked to the development of the filibranch gill and that the wider valve opening possibly would facilitate disposal of pseudofeces. These forms, which represent the earliest autolamellibranch bivalves, are now included in the Lower Ordovician to Lower Devonian superfamily

199

200

 . .  Cardiolarioidea (Cope 2000). Cardiolaria is the only form yet described from the Lower Ordovician and persists into the Middle Ordovician, where three new genera appear; of these four genera, three survive into the Upper Ordovician. Pojeta (1971, 1978), Pojeta and Gilbert-Tomlinson (1977), Tunnicliff (1982), and Cope (1996a) assigned (or assigned with a query) various bivalves to the genus Deceptrix Fuchs. Cope (1997b) showed that Deceptrix was a SilurianDevonian cardiolarioid and that the Ordovician forms assigned to it were praenuculids (nuculoids).

Trigonioids Trigonioids are distinctive bivalves with a prismatonacreous shell that became most important in the Mesozoic. Throughout their geologic history they appear to have lived successfully in high-energy, coarsegrained sands. The earliest, Noradonta, has dentition linking it to Cardiolaria, and from this the other Lower Ordovician genus, Tromelinodonta, can be derived. The latter is restricted to the Lower Ordovician but gave rise to the Middle Ordovician to Silurian genus Lyrodesma. Noradonta persisted into the Middle Ordovician in Australia, where an endemic genus Brachilyrodesma also occurs. In the Upper Ordovician, in addition to the almost cosmopolitan Lyrodesma, the elongate Pseudarca occurs; the latter is known from both Brittany and Wales (figure 20.2).

Heteroconchs Heteroconchs are bivalves that have their hinge teeth divided into cardinal, radiating from beneath the umbones, and lateral teeth, parallel to the dorsal margin of the shell. They may have originated from ancestors such as the cardiolarioids by fusion of the posterior teeth of the latter, already presaged in the cardiolarioid genus Eritropis, and lengthening of the anterior teeth. They include forms with a variety of shell structures including prismato-nacreous in some of the earlier forms and crossed-lamellar in many of the later ones; heteroconchs are the dominant group in Recent seas. A single genus, Babinka, is known from the late Tremadocian (TS.1d), but by the succeeding early Arenig (TS.2a) there is a wide variety of styles of dentition in the additional seven genera. In TS.2b is an unnamed genus, listed as “Actinodonta,”

that refers to French forms (Barrois 1891; Babin 1966) identified with the Silurian genus Actinodonta. The majority of these heteroconchs have ventrally radiating cardinal teeth and can be assigned to the family Cycloconchidae. The Glyptarcoidea have teeth radiating in the opposite direction and are close to the origin of pteriomorphians; this was either from the heteroconchs or directly from the ancestral cardiolarioid stock. Glyptarca itself has a dentition remarkably similar to that of Cardiolaria, as noted by Cope (1995). Cope (2000) retained the glyptarcoids in the heteroconchs, but the cladistic analysis of Carter et al. (2000) suggested to them that they were pteriomorphs. It is noteworthy that all the Lower Ordovician heteroconchs are from high latitudes. In the Middle Ordovician there are 12 genera of heteroconchs. Of these, 6 genera are new and 6 others are holdover genera. Although some species are remarkably common in certain faunas, others are much rarer, and the Spanish Middle Ordovician Ananterodonta is so far known from only a single specimen. There was now a low-latitude heteroconch, the Australian Copidens, which appears identical to the form figured by Guo (1988) as Zadimerodia from East Yunnan, China. Heteroconchs seem predominantly characteristic of high to intermediate latitudes and, as shown by Cope (2002), completely dominate the faunas at these latitudes. Heteroconchs are reduced to six genera in the Upper Ordovician, of which three are new (figure 20.3). Upper Ordovician faunas are well known from low-latitude areas where heteroconchs are not to be expected in great numbers. Of the six recorded genera, three are from low-latitude areas; the three high-latitude genus are all holdover genera. Because there has been little work on high-latitude Upper Ordovician bivalve faunas, the depletion in heteroconchs in the Upper Ordovician may be more apparent than real.

Anomalodesmatans Anomalodesmatans are burrowing bivalves with prismato-nacreous shells; throughout their geologic history they have constituted only a minor part of bivalve faunas. Dentition is frequently absent, and accurate valve articulation is provided by a well-located ligament; some forms have internal calcareous struc-

Tremadoc

Ibexian

1d 1c

Paramytilarca Anomalocoelia Praeanomalodonta Ambonychiopsis Leconychia Notonychia

Pteronychia

Glyptonychia

Cleionychia

Anomalodonta

Psilonychia

Warburgia Heikeia Thorslundia Ptychodesma Ambonychia

Vanuxemia

Ambonychioidea (pars)

Falcatodonta

2a

Pharcidoconcha

2b

Babinka

Arenig

ORDOVICIAN

2c

Cyrtodonta Cyrtodontula

Whiterockian Llanvirn

DARRIWILIAN

MIDDLE ORDOVICIAN

4a

3a

TREMADOCIAN

LOWER

4b

3b

472 Ma

489 Ma

4c

Cyrtodontoidea

Breviorthonota Ortonella

Paracyclas

5b 5a

460.5 Ma

Anomalodesmata

Rhytimya Orthonota Cymatonota Cuneamya "Ceromyopsis"

5c

Cycloconcha Neofordilla

Ashgill

5d Carotidens

UPPER

6b 6a

Mohawkian Caradoc

ORDOVICIAN

Cincinnatian

6c

Heteroconchia

Redonia Coxiconcha Moridunia Dulcineaia Taselasmodum Ananterodonta Carminodonta Famatinodonta Fortowensia Camnantia "Actinodonta" Celtoconcha Copidens Glyptarca Hemiprionodonta Arenigomya Sphenosolen

443 Ma

Pteriomorphia (pars)

TIME SLICES

North American & British series

GLOBAL SERIES & STAGES

Bivalve and Rostroconch Mollusks

1b 1a

FIGURE 20.3. Range chart of the genera of Ordovician Bivalvia belonging to the Heteroconchia, Anomalodesmata, and Pteriomorphia (part). In most cases the actual range within a time slice is not known, and the range is taken to the nearest upper and lower time-slice boundaries. Note that Carotidens is a pterioid and should appear in figure 20.4. For identification of higher level taxa in subheadings, see figure 20.2.

tures that may have helped valve articulation. Finescale tuberculation of the shell is another characteristic of the anomalodesmatans. The earliest genus and sole Lower Ordovician form is Arenigomya, from South Wales (Avalonia). There are two new Middle Ordovician genera, while in the Upper Ordovician the number of genera has risen to seven, six of which are new. All the Middle and Upper Ordovician genera are from low-latitude areas. The origins of the anom-

alodesmatans are obscure, but they may have been derived from early heteroconchs by loss of teeth.

Pteriomorphians The pteriomorphians are the most diverse groups of Ordovician bivalves; most have a three-layered shell with a calcitic outer layer and have multiple ligament insertions. They are frequently byssate. The

201

202

 . .  earliest is a Goniophora (or possibly the closely related Goniphorina; see Sánchez 1997). Goniophora and Goniphorina are modiolopsids, a group of predominantly infaunal bivalves that have been frequently classified in the past with the heteroconchs but shown recently (Carter and Seed 1998) to include forms with multiple insertions of nonparivincular ligaments, the latter being a characteristic of many pteriomorphians. Confusion also resulted from the linking of modiolopsids with the similarly shaped but later modiomorphids; the latter are now regarded as anomalodesmatans (Fang and Morris 1997). Modiolopsids became well diversified in the Lower Ordovician with two genera known from the upper Tremadocian of Australia and six genera known from the Arenig, including Goniophora; this form (Cope 1996b) may, like the Argentinian form discussed earlier, be a Goniophorina. Both are recorded as Goniophora on figure 20.4. It thus seems that this group was dispersed around the Gondwanan continent. Middle Ordovician modiolopsids were at a similar level of diversity with five recorded genera, three of which were holdover genera and two new genera, one from Australia and one from Wales. The dramatic change came within the Upper Ordovician whence 21 genera of modiolopsids have been recorded; this explosive increase occurred at low latitudes, not only on carbonate platforms such as on Laurentia and Baltica but also in Siberia and Kazakhstania. All these paleocontinents produced endemic modiolopsids. Of these 21 genera 15 are new, four are holdover genera, and two genera, Parallelodus and Eurymya, are known from the Lower Ordovician but not the Middle. Two genera (Corallidomus and Semicorallidomus) developed the habit of excavating living space (crypts) within stromatoporoids or bryozoans. With the Hirnantian regression and exposure of the carbonate platforms, both genera became extinct, and with them died the habit of boring among bivalves, which was not to be reevolved until the Late Jurassic. Cyrtodontids are pteriomorphians that have anterior teeth very like those of the glyptarcoids, but unlike the latter cyrtodontids have an edentulous subumbonal lacuna on the hinge plate; they also have multiple ligamental insertions, but the latter have not been recorded (?preserved) in some of the earlier examples. The earliest forms are from the late Tremadocian (TS.1d) of Australia, whence Cyrtodontula

and Pharcidoconcha were described by Pojeta and Gilbert-Tomlinson (1977); the latter genus also occurs in similar rocks in South China (Hsü and Ma 1948). Cytrtodontula also occurs in the early Arenig (TS.2a) of South Wales, where it is accompanied by Cyrtodonta and the strongly ribbed Falcatodonta (Cope 1996b). Cyrtodontids have a poor Middle Ordovician record; apart from four species of Cyrtodonta from Australia and two questionable records of Cyrtodontula from high latitudes in Spain and Morocco, the only other occurrence seems to be a late TS.4c occurrence of Vanuxemia from Kazakhstan (L. E. Popov pers. comm.). In the Upper Ordovician low-latitude carbonate shelves clearly provided a favorable environment for cyrtodonts, as in addition to Cyrtodonta, Cyrtodontula, and Vanuxemia there are five new genera and a large number of species. Pojeta (1971) recorded 116 species of Cyrtodonta—clearly an improbable total of real species but indicative of the group’s high diversity in the Upper Ordovician. Reduction of the anterior of the shell distinguishes the ambonychiids. The earliest are Cleionychia from the middle Arenig (TS.2c) of South Wales (Cope 1996b) and Notonychia, from the “Middle Arenig” (TS.3a) of Argentina (Sánchez 2001), possibly suggesting that this group originated at high latitudes, although it was clearly much more suited to low latitudes. The Middle Ordovician yields five genera, of which one is a holdover genus; three new genera are endemic to Australia and one to Argentina. In the Upper Ordovician there are 12 genera of ambonychiids, together with a large number of species, reflecting the low-latitude carbonate platforms as an ideal habitat for epifaunal bivalves. The end Ordovician regressions caused major extinctions on the carbonate platforms, and the ambonychiids were reduced to some five genera in the Silurian. Pterioideans are normally inequivalve bivalves with a flatter right valve than left, though, as they are descended from equivalve ancestors, the time of origin of this is not clear. The earliest is a single convex right valve of a Palaeopteria from the lower Arenig of South Wales (Cope 1996b). The only Middle Ordovician genus yet recorded is the Australian Denticelox. In the Upper Ordovician Palaeopteria reappears, and there are four new genera. Pterioideans became more important forms in the Upper Paleozoic.

Arenig

4a

2c

Incertae sedis Shaninopsis Shanina

Pholadomorpha Ischyrodonta Myodakryotus (L)

Prolobella (L)

Corallidomus

Aristerella

Orthodesma

Paraphtonia

Parallelodus Eurymya

Callodonta

Dipleurodonta

Modiolodon

Goniophora

Ahtioconcha

Modiolopsis Colpomya

Runnegaria

4b

Denticelox

4c

Paramodiola

Goniophorina Saffordia

Whiteavesia

Semicorallidomus

Pterinea Actinopterinea

Claudeonychia Palaeopteria

5b

Sphenolium

Modiolopsoidea + Limoidea (L)

Catamarcaia (A)

Whiterockian Llanvirn

DARRIWILIAN

2a

Tremadoc

Ibexian

1d

Colpantyx

2b Xestoconcha

MIDDLE ORDOVICIAN ORDOVICIAN

Maryonychia

5c

Allonychia

5d

3a

TREMADOCIAN

LOWER

Opisthoptera

Ashgill

Cincinnatian

6a

Mohawkian Caradoc

ORDOVICIAN UPPER

6b

3b

472 Ma

489 Ma

Ambonychioidea Pterioidea + (pars) Arcoidea (A)

6c

5a 460.5 Ma

Pteriomorphia (pars)

Matheria

443 Ma

TIME SLICES

North American & British series

GLOBAL SERIES & STAGES

Bivalve and Rostroconch Mollusks

1c 1b 1a

FIGURE 20.4. Range chart of the genera of Ordovician Bivalvia belonging to the Pteriomorphia (continued). In most cases the actual range within a time slice is not known, and the range is taken to the nearest upper and lower time-slice boundaries. Systematic position of the genera Shanina and Shaninopsis is uncertain, but both are probably pteriomorphians. Dotted lines signify ranges for which a genus must have survived but for which no record yet exists. Note that the pteroid Carotidens is incorrectly shown under the heteroconchs in figure 20.3. For identification of higher level taxa in subheadings, see figure 20.2. Abbreviations: (A) = Arcoidea; (B) = Limoidea.

The rarest Ordovician pteriomorphians are the arcoideans; they became much more important in the Upper Paleozoic and Mesozoic. They differ from other pteriomorphians in their two-layered shell structure. Their only Ordovician representative known hitherto is the genus Catamarcaia from the “Middle Arenig” (TS.3a) of Argentina (Sánchez and Babin 1993). As suggested by Cope (1997a, 1997b)

and Ratter and Cope (1998), Catamarcaia could be derived from a glyptarcoid heteroconch. Collection failure is clearly responsible for the disjunct record of the early arcoids; the next example, Alytodonta, is known from a single Lower Llandovery specimen. Catamarcaia cannot be accommodated into the family Frejidae, which includes all Silurian arcoids, so Cope (2000) proposed the family Catamarcaiadae.

203

204

 . .  Ratter and Cope (1998) suggested that the exceptional rarity of arcoids was due to their preference for close inshore environments, which are not often preserved. Two Upper Ordovician genera, Myodakryotus Tunnicliff 1987 and Prolobella Ulrich 1894 (TS.5b–c), are probably the earliest representatives of another group of pteriomorphians that became much more important later, the limoids. They were probably derived from the cyrtodonts and like them, but unlike later limoids, are dimyarian. ■

Rostroconchia

The Rostroconchia were the last class of the phylum Mollusca to be recognized. Although rostroconchs had been described since the early years of the nineteenth century, there was no recognition that they belonged to a single molluscan class. Much of the earlier literature had recognized them as members of three distinct groups of fossils: the ribeirioids, the eopteriids, and the conocardioids, based on the genera Ribeiria Sharpe 1853, Eopteria Billings 1865, and Conocardium Bronn 1835, respectively. The members of the last group were regarded as bivalves, usually allied to the cardioids, although by the time the bivalve volume of the Treatise (Branson et al. 1969) was published, they were treated as of uncertain bivalve subclass. The Treatise assigned Eopteria to the subclass Cryptodonta of the Bivalvia (Newell and LaRocque 1969), but both Kobayashi (1933, 1954) and Morris (1967) placed it with the ribeirioids; Pojeta (1971) however, preferred to group it with the conocardioids. Following Schubert and Waagen (1904), the ribeirioids were treated as arthropods, and this was followed by Kobayashi (1933) and most other earlier-twentieth-century workers. However, each of the three groups mentioned earlier has a calcareous shell bearing growth increments, developed from a protoconch, together with muscle scars, which also show growth increments. Such features are molluscan rather than arthropodan, and their recognition led to the concept of the class Rostroconchia by Pojeta et al. (1972) as a new class of bivalved mollusks, developed from a univalved protoconch. A major description and review of the class was published by Pojeta and Runnegar (1976) and a review of its geographic distribution by Pojeta (1979).

There are, however, differences between the numbers of species recorded herein and those produced by Pojeta (1979). There are several reasons for this. First, Pojeta’s (1979) numbers of species are inflated by inclusion of nomina nuda, the counting of subspecies as species, and the inclusion—as separate species—of species that were treated as synonymous by Pojeta and Runnegar (1976). In addition, Pojeta recorded as Middle Ordovician many species that would now be regarded as Upper Ordovician. Thus, from the high number of species recorded in the Lower Ordovician, there is a dramatic drop in the Middle Ordovician, followed by something of a recovery in the Upper Ordovician. Analysis of the distribution of species of rostroconchs shows that in the Early Ordovician there were 54 valid species distributed among 17 genera. This number has certainly been inflated by the records from two localities, particularly Manchuria, whence Kobayashi (1933) described many largely endemic genera and species. The other locality is in central Australia, whence Pojeta et al. (1977) described several endemic genera. The greatest generic diversification is evident in the Tremadocian, with no fewer than 16 genera, of which 9 are confined to that stage. There are some 21 Tremadocian species (figure 20.5). In the Middle Ordovician 10 of the 15 species recorded belong to genera that have survived from the Lower Ordovician. Three species belong to genera making their first appearance and that range into the Upper Ordovician; the remaining two species belong to genera restricted to the Middle Ordovician (figure 20.5). The Upper Ordovician shows a considerable increase in rostroconch diversity compared with the Middle Ordovician; although the number of genera has been reduced to seven, these contain 38 species. No new genera appear in the Late Ordovician, and of these, four survive into the latest Ordovician (TS.6c). Of the 38 species, some 26 are found on the Laurentian carbonate platform, and a further 7 species are known from the Baltoscandian carbonate platform. The remaining 5 species include 1 from lowlatitude Gondwanan carbonates (Tasmania), 1 from midlatitude carbonates (Kazakhstan), 2 from Bohemian clastic rocks, and 1 from Avalonian clastics. There are no substantiated records of rostroconchs from uppermost Ordovician (Hirnantian) rocks; this

443 Ma

TIME SLICES

North American & British series

GLOBAL SERIES & STAGES

Bivalve and Rostroconch Mollusks

Ribeirioida

Ischyrinioida

Ashgill

(1)

Talacastella (1)

(1)

(2)

Myocaris (1)

(2)

4a

Technophorus

(1) Pinnocaris

(4) Ribeiria

Whiterockian Llanvirn

DARRIWILIAN

(1)

?

Wanwanoidea (1)

Wanwanella (3)

Pseudotechnophorus (1)

Euchasma (1)

(9)

(8)

Eopteria

Tolmachovia (6)

Pauropegma (1)

(1)

(11)

Anisotechnophorus (1)

Eoischyrinia

1a

Aptoptopegma (1)

1b

Wanwania (2)

1c

Ptychopegma

Tremadoc

Ibexian

1d

(5)

2a

?

2b

(1)

(1)

ORDOVICIAN

2c

Ribeirina

Arenig

MIDDLE ORDOVICIAN

4b

3a

TREMADOCIAN

LOWER

4c

(2)

(2)

5b

3b

472 Ma

489 Ma

Hippocardia (6)

(8) Bransonia

(3)

(1)

5c

(13)

(5)

5d

Ischyrinia

UPPER

6b 6a

Mohawkian Caradoc

ORDOVICIAN

Cincinnatian

6c

5a 460.5 Ma

Conocardioida

FIGURE 20.5. Range chart of the genera of Ordovician rostroconchs. Information on ranges is for the most part accurate to one time slice, and ranges are taken to the nearest time-slice boundary unless clear evidence exists to the contrary. Numbers on each line indicate the number of valid species recognized for Lower, Middle, or Upper Ordovician (as appropriate). Dotted lines signify ranges for which a genus must have survived but for which no record yet exists. For identification of higher level taxa in subheadings, see figure 20.2.

must be due to collection failure, as rostroconchs continue through to the end of the Permian Period. Several conclusions may be drawn from examining these records of distribution of the rostroconchs. Generic diversity decreased throughout the Ordovician, and the number of monotypic genera decreased from six in the Early Ordovician to two in the Mid Ordovician; there are none in the Late Ordovician. Specific diversity decreased sharply from Early to Mid

Ordovician, but then increased again in the Late Ordovician but failed to regain the level of Early Ordovician diversity (figure 20.5).

Ribeirioids Of the order Ribeirioida, the Ribeiriidae, with a long Cambrian history, are more important in the Early Ordovician than later and are unknown from

205

206

 . .  the Hirnantian (latest Ordovician, TS.6c). Ribeiria itself showed a decline through the Ordovician and is probably represented by a single species in the Late Ordovician, but Pinnocaris, which appeared in the latest Cambrian and has so far not been recorded from the earliest Ordovician, reappeared in the mid part of the Early Ordovician and reached its maximum diversity (three species) in the Late Ordovician. The Technophoridae, with 13 species in the later part of the Early Ordovician, declined to 5 species in the Mid Ordovician and then increased to 13 species (all species of Technophorus) in the Late Ordovician. They, too, are unknown from the Hirnantian stage (latest TS.6c) but, unlike other ribeirioids, survive into the Early Silurian (Zhang 1984).

Ischyrinioids The monotypic order Ischyrinioida is restricted to the Ordovician. The family Ischyriniidae appears in the earliest Ordovician (TS.1a) with two species belonging to two genera. There is then a major break in the record of the family, as there are no records of it from the end of 1b to 4b in the later part of the Mid Ordovician, whence there is one record of an unidentified species of Ischyrinia. The genus is represented by five species in the Late Ordovician, though no ischyrinioids are recorded from the latest Ordovician (TS.6c).

Conocardioids The order Conocardioida is well represented from the earliest Ordovician. The monotypic superfamily Eopterioidea includes the family Eopteriidae, which includes 21 Lower Ordovician species belonging to four genera. One of these, Eopteria, is represented by one species in the Middle Ordovician, but there is only one other Middle Ordovician eopteriid, the South American monotypic genus Talacastella. Eopteriids are known from a single early Late Ordovician (TS.5a) species of Eopteria from North America and a single specimen from the mid Ashgill (mid-TS.6c) of Kazkhstan (Popov et al. 2003). The superfamily Conocardioidea has two families with representatives in the Ordovician. The Bransoniidae is represented by the genus Bransonia. This has a very disjunct range; a single species appeared in the early Tremado-

cian (TS.1a) of Australia, and the genus is then unknown until the Late Ordovician (TS.5a–6b), from which eight further species have been recorded. The Hippocardiidae appeared in the latest Mid Ordovician (TS.4c) with two species and in the Late Ordovician (TS.5a–6b) are represented by six further species. Neither Bransonia nor Hippocardia has been recorded from the latest Late Ordovician (latest TS.6c), but both are known from Silurian and later rocks. ■

Concluding Remarks

In summary, it appears that there were two phases to the Ordovician diversification of the Bivalvia. Knowledge of the first in the Lower Ordovician is hampered by the scarcity of faunas of that age, particularly from the Tremadocian. This initial explosive radiation, which may have followed on the evolution of the feeding gill, produced no fewer than 18 families representing most bivalve groups. The confinement of bivalves to Gondwana and closely bordering areas produced a steady increase in diversity through the Middle Ordovician. The irregular shape of the diversity curves reflects major publications (which in turn depend on individual faunas being found). This explains the peaks in TS.2a (Babin 1982; Cope 1996b) and in TS.4a (Pojeta and Gilbert-Tomlinson 1977; Babin and Gutiérrez-Marco 1991; Cope 1999). Other faunas may in time fill these gaps; for instance, a new fauna from West Yunnan, China (Fang and Cope in press), would raise the total generic diversity for TS.3a from 21 to 31, providing close to a linear increase from TS.2a to 4a. When bivalves managed to migrate to low-latitude carbonate platform areas, in the earliest Late Ordovician, there was major further diversification, especially within the pteriomorphians and the anomalodesmatans (see figure 20.6A), but heteroconchs were apparently markedly reduced; this latter effect may be the result of poor knowledge of high-latitude Late Ordovician faunas. The last diversification followed on from adoption of semi-infaunal and epifaunal habits and resulted in a dramatic increase in both generic and specific diversity, especially within pteriomorphians. The latest Ordovician eustatic regression, resulting from the growth of continental ice sheets, produced a major extinction among Ordovician bivalves, partic-

Bivalve and Rostroconch Mollusks

A

B

Group Diversity

Bivalve Faunal Diversity

Time Slices

Time Slices

C

D Origination

Time Slices

Per Capita Rate

Origination/Extinction

Time Slices

E Per Capita Rate

Extinction

Time Slices

Faunal Turnover

Per Capita Rate

F

Time Slices

FIGURE 20.6. Bivalve diversity measures. Time slices are accorded equal duration, apart from TS.1a–d, for which data are so sparse that those for TS.1a and 1b have been combined, as have those for TS.1c and 1d. Figures 20.6A and 20.6B should be compared. Note that the TS.4b diversity peak (figure 20.6B) is accounted for by equal proportions of the three bivalve group combinations (figure 20.6A), while the TS.5b–6b peak (figure 20.6B) is accounted for solely by the dramatic increase in Pteriomorphia and Anomalodesmata (figure 20.6A).

ularly among the pteriomorphian groups that had exploited the newly available ecologic niches on the low-latitude carbonate platforms. It is clear that the record of rostroconchs is deficient in several parts of their range; this is exemplified by the disjunct ranges of several taxa and the lack of any material from the Hirnantian (latest TS.6c). Clearly rostroconchs were profoundly affected by the Hirnantian regression; one species of ribeirioid is known from the Early Silurian and, surprisingly, appears to be the only record of a rostroconch from the Llandovery Series. The Conocardioidea, however, survived later into the Silurian and are recorded from the Wenlock Series upward to the end of the Paleozoic.

Pojeta (1979) suggested that an Ordovician decline in rostroconchs was correlated with bivalve diversification; he maintained that in competition between two infaunal groups, the superior burrowing powers of the Bivalvia gave them a decided advantage. Examination of the records of the rostroconchs (figure 20.7) and bivalves (figure 20.6) shows, however, that this idea can no longer be countenanced. First, the initial Early Ordovician bivalve diversification took place around the Gondwanan margins, where there were few rostroconchs. Second, the early Mid Ordovician rostroconch decline was largely extra-Gondwanan, whereas bivalves were still almost exclusively Gondwanan. Third, the Late Ordovician increase in rostroconch diversity, particularly on the

207

208

 . .  Rostroconch Diversity

Time Slices

FIGURE 20.7. Rostroconch diversity. Time slices have been accorded equal duration. Since the decline is rostroconch genera in TS.1, generic diversity remained remarkably consistent (see figure 20.5).

Laurentian shelves, coincided with the arrival of bivalves on these shelves where the latter rapidly diversified. Rostroconchs were infaunal, but the Late Ordovician bivalve diversity increase was predominantly (though not exclusively) among epifaunal groups. Competition between bivalves and rostro-

conchs was clearly not a significant factor in rostroconch distribution through most of the Ordovician; the first time there could have been such competition was in the Late Ordovician, when significant numbers of bivalves and rostroconchs coexisted. Both classes were profoundly affected by the latest Ordovician glacially induced regression; bivalves largely recovered from this, but rostroconchs were reduced to one genus of ribeirioids, which survived through to the Llandovery Series, and one superfamily of conocardioids, which persisted through to the end of the Permian Period.

 I am grateful for the assistance of L. C. Norton (National Museum of Wales) in drafting figures 20.1– 20.5 and A. P. Rogers (Cardiff University) for help with figures 20.6–20.7.

21

Nautiloid Cephalopods Robert C. Frey, Matilde S. Beresi, David H. Evans, Alan H. King, and Ian G. Percival

T

he Ordovician Period marks the time of the great radiation of nautiloids, which proliferated from a single order at the beginning of the Ordovician (489 Ma) to a maximum of at least nine orders by the early Late Ordovician (455 Ma). It is also in the Ordovician that the greatest diversity of shell form and structures occurs in these externally shelled cephalopods (Flower 1976; Holland 1987; Teichert 1988). Most of these various shell architectures were new solutions to the problem of buoyancy control in these mobile mollusks (Crick 1988). Ordovician nautiloids are represented in strata exposed from the Arctic Circle to Tasmania and have been identified from lithofacies indicative of diverse marine environments. In the Ordovician, however, nautiloids are most abundant, display maximum diversity, and reach their greatest size in shallow marine carbonate platform facies deposited under tropical or subtropical climatic conditions (Flower 1976). ■

Status of Nautiloid Systematics

Although much has been achieved over the past century as far as the study of systematics of Ordovician nautiloids is concerned (Flower 1976; Teichert 1988), a number of issues remain to be dealt with, especially the validity of many published species. The tendency for some workers in the past to split nautiloids excessively at the species level was due to a lack

of established criteria to define and differentiate taxa, and description of new species based on incomplete material. This often resulted in the application of several species names to different portions of the shell of the same organism. Another complication is the possibility of sexual dimorphism in fossil nautiloids, leading to the potential for different sexes of the same species to be identified as separate species. As a result of these unresolved difficulties at the species level, this review of nautiloid diversity in the Ordovician utilizes genus-level taxa, which are much better established. The higher-level taxonomic classification used here is modified from the phylogenetic diagrams of Flower (1988) and Wade (1988). This includes the recognition of the Pseudorthocerida as an order distinct from the Orthocerida; the merging of the Barrandeocerida with the Tarphycerida; and the recognition of an additional new order, the Dissidocerida, proposed by Zhuravleva (1994) since publication of the papers just cited. Family-level taxa include those listed in the compilation by King (1993). ■

Regional Distribution

Knowledge of the ranges and distribution of Ordovician nautiloid genera in specific geographic regions is highly variable. Ordovician nautiloid faunas have been extensively studied and well documented for North America (Flower 1976); Baltic Europe 209

210

 .   . (Flower 1976; King 1999), Bohemia (Marek 1999), Russia and Siberia (Balashov 1962), and Australia (Percival in Webby et al. 2000). Chinese Ordovician nautiloid faunas, both for North China and Korea and for South China, are relatively well established (Chen 1995), although in much of the older literature (Kobayashi 1927; Yu 1930; Endo 1932) many Chinese forms were ascribed to established North American and Baltic taxa when at least some of these genera represent new taxa. Nautiloid faunas from classic Ordovician localities in the United Kingdom and other portions of the Avalonian microcontinent are inadequately known because of poor preservation of the nondescript longiconic nautiloids that dominate faunas in these shale and mudstone strata. Current work in South America, especially from Ordovician sections in Argentina and Bolivia, suggest that diverse, well-preserved faunas are present that are yet to be documented properly. Study of these South American faunas is also complicated by the disparate geologic histories of the Argentine Precordillera and Gondwana portions of northern Argentina and Bolivia (Astini 1998; Benedetto 1998). Also relatively poorly understood are Ordovician nautiloid faunas from southern Europe, the Middle East, Central Asia, and Southeast Asia, which are known only from a limited number of systematic studies. ■

Measured Parameters

Nautiloid diversity patterns were tabulated using normalized diversity after Cooper (chapter 4). Also calculated were turnover rates per million years, which record the number of genera originating or ending in each time unit divided by its duration. Figures 21.1 and 21.2 graphically display normalized diversity and turnover rate data (both originations and extinctions) for nautiloid genera for each of the identified time slices (TS ) recognized here as subdividing the Ordovician Period. ■

FIGURE 21.1. Normalized diversity of nautiloid genera through Ordovician time.

Summary Trends in Nautiloid Diversity

The interpretation of general trends in nautiloid diversity in the Ordovician based on the faunas studied here is limited by a number of factors. The authors have compiled available data on the geologic

FIGURE 21.2. Originations and extinctions of nautiloid genera through Ordovician time.

and geographic occurrence of Ordovician nautiloid order-, family-, and genus-rank taxa from North America (Frey), the Baltic region (King), Avalonia (Evans), South America (Beresi), and Australia (Percival). The absence of data for significant geographic regions, especially the Siberian Platform and North and South China, calls into question interpretations of global trends in nautiloid diversity. Diversity trends, especially for mobile continent-bearing plates such as Avalonia, the Baltic region, and the Argentine Precordillera, appear to have been strongly influenced by regional geologic and biogeographic events. These factors and the effects of incomplete data for some

Nautiloid Cephalopods regions and time periods (South America: entire Ordovician; North America: TS.4b–c) limit our confidence in the broad-scale diversity trends identified here. These uncertainties and limitations being stated, some general trends in nautiloid diversity in the Ordovician can be identified from these data. In support of the trends identified here, there is general agreement between these tabulations and standing generic diversities presented by Crick (1990), with the possible exception of the Late Ordovician (TS.6a–b).

Early Ordovician (TS.1a–2c) Information on nautiloid diversity in the Early Ordovician is based primarily on faunas described from North America and from Western Australia. Both of these areas were sites of tropical-subtropical platform carbonate deposition in the Early Ordovician. Studied faunas in TS.1 (Tremadocian) mark the first great expansion of the cephalopods with the evolution and diversification of a number of small longicones and cyrtocones belonging to the order Ellesmerocerida. These earliest Ordovician faunas are dominated by members of the family Ellesmoceratidae. Early representatives of the Baltoceratidae, Bassleroceratidae, and Protocycloceratidae and the orders Endocerida and Tarphycerida evolve later during TS.1b–d (Flower 1964, 1976). Faunas representative of TS.2a–c (early Arenig equivalents) are enriched by the diversification of the ellesmerocerid families Baltoceratidae, Cyclostomiceratidae, and Protocyloceratidae. The orders Endocerida (breviconic Piloceratidae, longiconic Proterocameroceratidae) and the coiled Tarphycerida (Tarphyceratidae) underwent rapid expansion, and the first members of the orders Orthocerida, Dissidocerida, and Actinocerida occur. Twenty-two new genera are added during TS.2a (figure 21.2), leading to an Early Ordovician maximum of diversity in TS.2b. Australasian faunas include a diverse group of endemic ellesmerocerids, endocerids, and the oldest members of the order Actinocerida, plus more cosmopolitan tarphycerid genera (Teichert and Glenister 1954). The end of the Early Ordovician (TS.2c) exhibits a slight decrease in diversity resulting from the extinction of a number of ellesmerocerid genera, primarily members of the Ellesmeroceratidae that

had been persistent for much of this time interval (figures 21.1, 21.2).

Mid Ordovician (TS.3a–4c) According to Flower (1976:528), the Mid Ordovician marks “a definitive point in the evolution of the Nautiloidea” with the expansion of the orders Endocerida, Actinocerida, and Orthocerida and the first occurrences of the orders Pseudorthocerida, Oncocerida, and Discosorida. The Tarphyceratidae decline dramatically, except for relic forms that persist in the Baltic region. The Trocholitidae and Lituitidae diversify as the Tarphyceratidae decline. Documented faunas fall into two distinctive biogeographic groups (Flower 1976): 1. An equatorial carbonate platform fauna represented by faunas from North America, North China, Korea, and Australia, which is distinguished by a collection of robust endocerids with complex endosiphuncular structures (Allotrioceratidae, Endoceratidae, Padunoceratidae), large actinocerids belonging to the Wutinoceratidae, a variety of rather simple orthocerids, robust coiled tarphycerids of the Trocholitidae (Litoceras, Plectolites), and primitive oncocerids. 2. A higher-latitude, Gondwanan fauna represented by genera from both clastic and carbonate strata in the Baltic region, South China, southern Europe, Avalonia, and portions of South America. This fauna is distinguished by specialized ellesmerocerids with greatly thickened connecting rings (Bathmoceratidae), a number of slender longiconic endocerids (Dideroceras and Suecoceras), the cyrtoconic endocerid Cyrtendoceras, true Orthoceras and its close relatives (Orthoceratidae), the long-ranging, widespread coiled trocholitid genus Discoceras, and numerous members of the variably uncoiled Lituitidae. An undescribed Gondwanan fauna from southern Bolivia also includes large endocerids (Padunoceratidae) and actinocerids that appear to be assignable to the Georginidae (Flower 1988). Nautiloid diversity reaches a second Ordovician maximum in TS.4b, primarily owing to the presence of a well-documented diverse Baltic fauna (Sweet 1958; King 1999). Many of the Mid Ordovician genera from both paleogeographic groups appear to have suffered extinction at the end of TS.4c (figure 21.1).

211

212

 .   . These extinctions may in part be the result of a global sea level drop at this time (Chen 1991) that is often overprinted by a post-Darriwilian unconformity, as is the case in North America (Flower 1970).

Late Ordovician ( TS.5a–6c) The early Late Ordovician (TS.5—Caradoc equivalent) is marked by continued diversification of the Orthocerida and the Pseudorthocerida and the rapid expansion of the Oncocerida. New families added include large actinocerids belonging to the Actinoceratidae, the Gonioceratidae, and the Armenoceratidae, plus large tarphycerids with cyrtochoanitic siphuncles (Barrandeoceratidae, Apsidoceratidae). Two new orders, the Ascocerida, with their curious truncated mature phragmocones, and the cyrtoconic and breviconic Discosorida, diversify during this time interval. However, diversity within the Endocerida, Trocholitidae, and Lituitidae declines significantly. The Ellesmerocerida, with the exception of a few baltoceratid genera and the peculiar curved brevicone Cyrtocerina, have disappeared. Some of the geographic distinctions between nautiloid faunas disappear as Baltica and Avalonia approached Laurentia (Mac Niocall et al. 1997). A common fauna developed in association with carbonate platform facies distributed across a vast “tropical realm” extending from Greenland to Mexico (Flower 1976) and spreading eastward into now adjacent portions of the Baltic region (Strand 1933) and approaching portions of the Avalonia microcontinent (Evans 1993). This diverse “Arctic Ordovician Fauna” included often gigantic individuals belonging to the Endocerida, Dissidocerida (Narthecoceratidae), Actinocerida, Tarphycerida (Apsidoceratidae), Ascocerida, Oncocerida (Diestoceratidae), and Discosorida (Cyrtogomphoceratidae, Westonoceratidae) (Foerste 1929). Elsewhere, South American faunas appear to decline dramatically as a result of drift into higher latitudes coupled with the onset of continental glaciation in Gondwana at the end of the Ordovician. Rare longiconic orthocerids occur associated with a “Hirnantia” shelly fauna in marine mudstone facies (Sánchez et al. 1991). Australian faunas persisted in carbonate platform facies centered in Tasmania and consisted of curious mixtures of cosmopolitan genera (Beloitoceras, Gorbyoceras, Discoceras) associated with

distinctly Tasmanian taxa belonging to the endemic Gouldoceratidae (order Discosorida) (Stait 1988). The later Late Ordovician (TS.6—Ashgill equivalent) demonstrates a continuation of diversity trends set in earlier in the Late Ordovician and marks the maximum in terms of generic diversity for nautiloids in the Ordovician. There is a major extinction of nautiloid genera and families at the end of the Ordovician (figure 21.2), associated with Late Ordovician glaciation events (chapter 9). A marked diversity decline occurs between TS.6b and 6c (figure 21.1). However, relic Ordovician faunas of endocerids, orthocerids, dissidocerids, actinocerids, and discosorids continued on into the Early Silurian (Rhuddanian-Aeronian), evidently finding refugia in carbonate platform habitats in Laurentia (North America, Greenland, the Arctic archipelago) (Frey and Holland 1995). ■

Comparisons with Published Biodiversity Analyses

Earlier we noted general (though not exact) agreement between our data and standing generic diversities presented by Crick (1990), with the possible exception of the Late Ordovician (TS.6a–b). A subsequent analysis by Sepkoski (1995) of a number of the major Ordovician clade groups includes a graphic representation of nautiloid radiation and extinction within the period, which in places also differs in detail from the generic diversity trends depicted here (figure 21.1). For example, for the Arenig interval, Sepkoski counted 125 genera, and Crick listed 178 genera. Our compilation for the Arenig (TS.2a–b, 3a–b, 4a) totaled 105 genera, an estimate that seems to be generally in line with Sepkoski’s data (and to a lesser extent Crick’s tabulation), especially given the absence of generic information in our analysis from the Siberian Platform, former Soviet Central Asia, and North and South China. We recognize an Arenig biodiversity maximum near the end of the Lower Ordovician (TS.2b), resulting from a burst of evolutionary activity leading to the development of many new taxa that started in TS.2a. Data compiled for the Arenig interval in North America (by Frey, herein) included a block of new genera and species from unpublished manuscripts by the late Rousseau Flower, describing faunas from the El Paso region of Texas and New

Nautiloid Cephalopods Mexico and from Newfoundland. These data push the origins of many taxa, principally members of several endocerid families (especially the Piloceratidae), back to a point earlier in the Arenig rather than appearing toward the end of the interval as has been traditionally believed (reflected by Sepkoski’s maxima at the Arenig-Llanvirn boundary). For the Llanvirn (TS.4b–c), Crick (1990) tabulated 131 genera, whereas our compilation listed only 75 genera. This interval marked the occurrence of our second maximum of 43 genera. Llanvirn nautiloids are best documented from two regions—the Baltic, which is covered in our compilation, and two Chinese blocks, neither of which are included here. Absence of these Chinese taxa could very well account for the evident discrepancy between our numbers for total diversity and those of Crick. Sepkoski (1995) recognized another maximum at the base of the Caradoc. We have a minor peak in diversity (37 genera) in roughly the same place. Crick (1990) recorded 135 genera from the entire Caradoc interval, not significantly different from our total of 125 genera. However, our compilation includes the Edenian (TS.5d) in the Caradoc, with its exceptionally diverse Laurentian Arctic Ordovician nautiloid fauna. Our compilation with regard to total Ashgill nautiloid genera was 65 (TS.6a–c). This is approximately half the total number of genera tabulated by Crick (1990) for his Ashgill interval (129 genera). According to Crick’s data, the most diverse Ashgill faunas were from North America (95 genera) and the Baltic region (56 genera). Both areas were included in our compilation but with significantly different results: North America (65 genera) and the Baltic region (37 genera). If TS.5d is added to the Ashgill interval (as probably was the case with Crick’s data), generic diversity climbs to 81 for North America and 39 for the Baltic, closing the gap somewhat but still leaving significant differences. As Ashgill faunas are well documented for both regions, these discrepancies are somewhat more difficult to explain. A further maximum was identified by Sepkoski in the late (but not latest) Ashgill. Although we recognize a maximum of diversity (47 genera) somewhat earlier in the Ashgill (TS.6a), this discrepancy probably relates to slight differences in the boundaries of the time intervals used in the respective compilations.

With regard to these apparent differences between the compilations of Ordovician nautiloid diversity by Sepkoski (1995) and trends based on the diversity calculations used here, there are a number of possible explanations. The most likely are as follows. First, Sepkoski’s compilation was based on worldwide data, whereas the present analysis was limited to North America, South America, the Baltic region, Avalonia, and Australia; significant faunas from the Siberian Platform, Central Asia, and both Chinese blocks are missing from our compilation. Second, differences in correlations used in these respective studies could have a significant effect on diversity per time interval tabulations. Finally, both Sepkoski’s and Crick’s tabulations are of simple generic diversity, whereas our compilation employed a “normalized” measure of generic diversity. Comparison of these two biodiversity measures for the sample time intervals indicates that our normalized generic diversity represents only two-thirds of the total number of genera present. However, it is important to note here that in our compilation of Ordovician nautiloid genera, some subjective discretion was used in tabulating generic diversities. Not included in our counts were a number of questionable occurrences; also eliminated were erroneous assignments of Ordovician forms to Silurian and Devonian genera. These include orthocerid genera such as Cycloceras, Dawsonoceras, Geisonoceras, Geisonocerina, Leurocycloceras, Michelinoceras, Protokionoceras, Spyroceras, and Virgoceras, as well as oncocerid genera such as Cyrtoceras, Gomphoceras, and Mixosiphonoceras. Another possible reason for apparently lower generic diversity involved the removal from our lists of a number of forms now generally agreed to be synonyms of other recognized genera. For example, the endocerid genera Cyclendoceras Grabau and Shimer, Foerstellites Kobayashi, and Nanno Clarke are all regarded as synonyms of Endoceras Hall. In conclusion, all the factors discussed here have played a role in causing the apparent differences between our compilation and those of previous authors including Crick (1990) and Sepkoski (1995). Each of these tabulations (including that presented here) represents a snapshot of Ordovician nautiloid diversity as it was known at the time of the compilation, each with its own set of limitations and factors that affect its accuracy and completeness.

213

22

Tube-Shaped Incertae Sedis John M. Malinky, Mark A. Wilson, Lars E. Holmer, and Hubert Lardeux

T

he hyoliths, cornulitids, coleoloids, sphenothallids, bryoniids, and tentaculitids included in this chapter represent unrelated, exclusively Paleozoic benthic and pelagic groups of organisms, with radial to bilateral symmetry, solitary (or rarely clustered), tube- or cone-shaped shells, and calcitic (possibly in some, originally aragonitic), phosphatic, and organicwalled shell preservation. The conical-shelled, operculate hyoliths are a moderately diverse Ordovician group, and though they did not reach the peak of abundance and diversity attained previously, in the Cambrian, they did become well diversified in higherlatitude cooler waters of the Mediterranean Province. The smaller groups comprising the mainly calcitic cone-shaped solitary coleoloids, the branched (compound) sphenothallids, and the phosphatic bryoniids with tubes and disklike attachments also have an earlier (Cambrian) history. This contrasts with the solitary conoidal, annulated shells of cornulitids (with characteristic longitudinal striae and vesicular wall) and tentaculitids (with a multilayered wall and internal septa), which do not appear until Late Ordovician, specifically mid Caradoc (= North American Mohawkian), time. ■

Hyoliths ( JMM)

The class Hyolitha Marek 1963 (phylum Mollusca) encompasses calcareous, operculate, conical-shelled 214

organisms, ranging from Early Cambrian to at least Mid Permian (Fisher 1962). The class includes two main groups, the order Hyolithida Syssoiev 1957 (Early Cambrian to Mid Permian), and the order Orthothecida Marek 1966 (Early Cambrian to Mid Devonian). The former is distinguished by the presence of a projection, or ligula, along the ventral rim (but see Kruse 1997) of the aperture and curvilinear structures known as helens, which protruded from the aperture even when the operculum was closed. These structures either may have functioned to stabilize the animal or may have permitted limited movement on the seafloor. The interior of the operculum possesses radially arranged structures known as clavicles, which occur as either a single pair or up to seven pairs in some taxa. The Orthothecida lacks the helens and ligula along the apertural rim, having instead a planar aperture or in some taxa an apertural rim with indentations, and the interior of the operculum lacks clavicles and other structures seen in the Hyolithida. The affinity of the Hyolitha remains a matter of controversy. Many early workers supported molluscan affinity, based largely on similarity in shell structure between hyoliths and other mollusks (Marek and Yochelson 1976). In contrast, other studies suggest separate phylum status (Runnegar et al. 1975). The issue of hyolith affinity remains unresolved, although herein hyoliths are regarded as mollusks.

Tube-Shaped Incertae Sedis Various aspects of hyolith ecology have long been controversial. A planktic mode of existence has been favored by many early workers owing to the supposed affinity of that group to modern pteropods (see Fisher 1962 for summary), whereas more recent investigations strongly suggest that these organisms likely were epifaunal (Marek and Yochelson 1976) or perhaps semi-infaunal. Hyoliths have long been interpreted as deposit-feeding organisms, an observation supported for the Orthothecida by the nature of the intestine (Marek and Yochelson 1976). More recently the Hyolithida have been reinterpreted as suspensionfeeding organisms with the aperture oriented toward oncoming currents (Marek et al. 1997). Hyoliths are found in all normal marine facies. Abundance and diversity attain a maximum in the Cambrian, followed by a major decline, perhaps due to increased competition from gastropods and other benthos. Only widely scattered, sporadic occurrences usually consisting of one or several individuals are known in post-Cambrian Paleozoic rocks. Kammer et al. (1986) considered the hyoliths to be a relict of the Cambrian Evolutionary Fauna (Sepkoski 1981a), with post-Cambrian occurrences largely confined to facies from stressed marine environments, where competition from other normal marine benthos would presumably have been less. This observation is supported by the abnormally high numbers of 76 hyoliths from an ooid shoal facies in the Mississippian of Iowa (Malinky and Sixt 1990) and several hundred from oxygen-deficient gray shale facies in the Midcontinent Pennsylvanian (Malinky and Mapes 1983). The Ordovician hyoliths discussed herein are an apparent exception, but high diversity and abundance in Sweden are due at least partly to a concentration of these fossils at a diastem (Dronov and Holmer 1999) and in Bohemia to unusually good circumstances of preservation (Marek 1967, 1989). Data on diversity of Ordovician hyoliths are given later in this chapter. One aspect of hyolith paleontology that has long been neglected is the recognition of geographic and stratigraphic biodiversity patterns within the group. Marek (1976) was the first worker to provide a modern synthesis that emphasized but was not restricted to Ordovician hyoliths from Bohemia (Czech Republic) and later added an updated version (Marek 1989). Major impediments to formulating accurate statements on hyolith biodiversity include uncertainty

about the morphological boundaries of the group, although this applies more to the Cambrian tubular fossils from Siberia and China than to any Ordovician specimens. Furthermore, with the exception of Marek (1976, 1989), whose works largely utilized recently collected material, workers were compelled to rely on museum collections assembled since the nineteenth century because in general hyoliths are such rare fossils. As a result, in many cases the stratigraphic level, and in some instances even the locality, cannot be identified with certainty (Malinky 2002). Usually the occurrences can be recorded only in terms of broader regional stratigraphic groupings. As a result, the number of taxa within a time slice cannot be determined with certainty; nor can they be related to particular environmental events. Hyolith species are rarely found in Ordovician sequences, and they characteristically constitute a lowdiversity component of a more diverse assemblage. Possibly this is a consequence of their low species abundances. Hyolith biodiversity is far more difficult to assess because of the overall paucity of specimens compared with other marine Ordovician invertebrates. A further complication arises from the quality of preservation of these fossils, which greatly limits how adequately some species can be identified. The skeletal mineralogy of the group has long been recognized as consisting of some unstable material (Holm 1893), possibly aragonite (Marek and Yochelson 1976), which would account for the overall poor preservation. The present summary focuses mainly on material from areas of central Europe (Bohemia) and Morocco and from Scandinavia, within limits of the Mediterranean and Baltic faunal provinces, respectively (see Marek 1976). The species from these two provincially distinct regions have been revised recently using modern taxonomic principles and the stratigraphic control is far better than that of material from elsewhere. The present contribution is a revised version of Marek’s (1976) study expanded to include recently restudied hyolith taxa (e.g., Malinky 1990, 2002). Marek’s (1976) recognition of separate Mediterranean and Baltic hyolith provinces is further reenforced by data presented herein (figure 22.1). Significantly, the geographic distribution of some taxa is now known with much greater precision than their temporal distribution, which is the main subject of this contribution. Many Ordovician hyolith species listed by

215

Cincinnatian Ashgill Harju

Gompholites Hexitheca Hyolithes Joachimilites Leolites Mediolites Nephrotheca* Nervolites Panitheca* Pauxillites Quadrotheca* Raitilites Recilites Solenotheca Sololites Stelterella Sulcavitus Trapezotheca*

Elegantilites Eumorpholites Gamalites

TIME SLICES

Bactrotheca* Brevitheca* Carinolithes Cavernolithes Chelsonella* Chimerolites Circotheca* Crispatella Decipilites Dilytes Dorsolinevitus

6c 6b

(1)

(1)

(1)

6a

Mohawkian Caradoc

5d

5c (5)

5b Viru

UPPER ORDOVICIAN

443

(1)

(1)

(3)

5a

Arenig

Volkhov

DARRIWILIAN Whiterockian Llanvirn Kunda

MIDDLE ORDOVICIAN

(5)

LOWER ORDOVICIAN

FIGURE 22.1. Stratigraphic distribution of hyolith genera, with number of species known for each genus given in each column, including species in open nomenclature. Distributions are shown only in those intervals in which genera can be associated with a specific province. For example, Hyolithes ranges into the Lower Devonian, but because the fossil record of that genus is incompletely known, it is uncertain whether it was still present in Baltica in the Late Ordovician or whether by that time it had migrated to the Mediterranean Province. Orthothecids are indicated by an asterisk next to the name of the genus. Note that for the Baltic Province the Tremadoc has been interpreted as having a more abbreviated scope than in traditional British usage, now equivalent to the Pakerort and Varangu stages combined, while the Latorp is now represented by the Hunneberg and Billingen stages combined (see figure 2.1).

GLOBAL SERIES & STAGES North American British and Baltic Units

 .   .

489 Ma

4c (1) (5)

4b

(1) (4)

(1)

4a

(1) (3)

(3)

(6)

(1) (2)

(2)

3b (2)

(2)

3a

(1)

(1) (3)

2c 2b

TREMADOCIAN Ibexian Tremadoc Tremadoc Latorp

216

2a 1d

(3)

Distribution and Diversity Patterns Ordovician hyoliths have a worldwide distribution, with occurrences reported from all continents except Antarctica, but except for Europe and North Africa, the number of species described is small. Sinclair’s (1946b) compilation consisted of about 25 species worldwide. That number has substantially increased since then primarily owing to the works of Marek cited herein.

Generic Patterns In terms of the two distinct biogeographic provinces Marek (1976) recognized that 23 hyolith

(9)

(4)

(12)

1c 1b 1a

Mediteranean Province

Sinclair (1946b) are in need of revision and therefore cannot be incorporated into the present biogeographic and biodiversity schemes.

(7)

(3)

(3)

Baltic Province

United States

genera were present in the Mediterranean Province and 8 genera in the Baltic Province, with only 2 genera, Carinolithes and Quadrotheca, common to both in the Ordovician (figure 22.1). Circotheca occurs in both provinces, but it is exclusively a Mediterranean form in the Ordovician and solely a Baltic element in the Cambrian (Berg-Madsen and Malinky 1999). Marek (1976) stated that Circotheca also occurs in Germany and Bolivia, but these occurrences await confirmation. Carinolithes also occurs in Britain as Hyolithus pennatuloides (Malinky unpubl.), along with Leolites and Recilites (Malinky 2003). Revision of the few specimens reported from North America indicates that only two genera are recognizable— Chelsonella (Lower Ordovician) and Solenotheca (Middle Ordovician)—and these have a Laurentian distribution. A marked endemicity is exhibited by the hyolith faunas of the Mediterranean and Baltic provinces.

Tube-Shaped Incertae Sedis This becomes even more evident with the incorporation of both newer hyolith taxa and taxa from other regions into the present paleogeographic/biodiversity summation. The few exceptions are Carinolithes, which occurs in both provinces but is represented by different species in each and has a much longer stratigraphic distribution in the Baltic Province (figure 22.1). Hyolithes is exclusively a Baltic form in the Ordovician but is found in Bohemia during the Silurian and Devonian. Quadrotheca occurs in both provinces, but at different times in the Ordovician. Ten hyolith genera are known in the Early Ordovician, with a significant overall increase to 23 in the Mid Ordovician. Nearly two-thirds of the Mid Ordovician fauna is composed of Mediterranean generic components, and most of the remainder of the fauna consists of Baltic genera. All of the 17 reported from the Late Ordovician belong to the Mediterranean Province. The significant loss of hyolith diversity from the Baltic Province may perhaps be linked to the progressive cross-latitude drift of Baltica though Ordovician time, with the lithospheric plate moving from higher South Polar paleolatitudes to lower latitudes by Late Ordovician time (Torsvik et al. 1990). The Mediterranean and Baltic provinces were only very weakly differentiated in the Early Ordovician (especially Tremadocian; TS.1a–d; see figures 22.1 and 22.2 and time slices in chapter 2), only becoming really well developed during the Mid Ordovician. Initially in the Tremadocian the hyolith generic diversity was low, comprising only two Mediterranean genera, Cavernolithes and Elegantilites from Morocco (Marek 1983), and the Baltic genera Carinolithes and Hexitheca from Sweden (Berg-Madsen and Malinky 1999; Malinky and Berg-Madsen 1999). One Lower Ordovician genus, Chelsonella, is recorded from Laurentia. In the Arenig (TS.2a–4a), a more diverse Mediterranean fauna comprising Bactrotheca, Cavernolithes, Circotheca, Elegantilites, Eumorpholites, Gamalites, Gompholites, Nephrotheca, Nervolites, Panitheca, and Pauxillites is recorded from Bohemia, with similar occurrences in Morocco in the same interval (Marek 1976, 1983, 1989). Many of the Mediterranean Arenig genera have extended ranges in Bohemia, occurring along with the new Llanvirn (TS.4b–c) components Dilytes, Leolites, and Recilites, resulting in a slightly higher generic diversity.

40 (38)

HYOLITHIDA

X

Mediterranean Province Baltic Province Laurentia

35

ORTHOTHECIDA

30

(29)

Mediterranean Province Baltic Province Laurentia

25

(25)

X

TOTAL SPECIES DIVERSITY: X 20

(19)

X

15 (12) (11)

10

X

(9)

(9)

X

(6)

(4) X

1a

X (3)

(2) (2)

0

(5)

(3)

(2) (1)

1b 1c 1d Tremadoc

Tremadoc

(1)

2a

(1)

2b

Latorp

Volkhov

(1)

Kunda

5c Caradoc

5a

5b

5d

6a 6b 6c Ashgill Harju

Viru

Whiterockian

TREMADOCIAN

LOWER ORDOVICIAN

(2)

4c 2c 3a 3b 4a 4b Arenig Llanvirn

Ibexian

489 Ma

(7)

(6)

(5)

5

X

(9)

(9)

(8)

Mohawkian

Cincinnatian

DARRIWILIAN MIDDLE ORDOVICIAN

472

460.5

UPPER ORDOVICIAN 443 Ma

FIGURE 22.2. Number of species through time. Species occurrences are given in terms of broad regional stratigraphic units rather than time slices because of imprecisely known stratigraphic data for most species. For the Mediterranean Province, Tremadoc refers to TS.1a–d; Arenig: TS.2a–4a; Llanvirn: TS.4b–c; Caradoc: TS.5a–c; and Ashgill: TS.5d–6c. For the Baltic Province, Tremadoc is TS.1a–b; Latorp: TS.1c–2c; Volkhov: TS.3a–4a; Kunda: TS.4a–b; Viru: TS.4c–5c; and Harju: TS.5c–6c.

The provincially distinct Baltic hyoliths become well differentiated only during the Mid Ordovician. Hexitheca disappears at the Lower Ordovician– Middle Ordovician boundary, whereas Carinolithes persists throughout the entire Mid Ordovician (with migration to the Mediterranean Province during the Mid Ordovician). Quadrotheca appears near the top of the Volkhov (TS.4a), whereas Dorsolinevitus, Hyolithes, Sulcavitus, and Trapezotheca appear slightly later though in the same time slice near the VolkhovKunda boundary. Crispatella first appears in the lower Kunda (TS.4b–c), and Stelterella is confined to the Viru (TS.4c). Solenotheca is the sole recognizable Mid Ordovician hyolith genus from Laurentia. In the Caradoc (TS.5a–d), all recognizable hyolith genera are confined to the Mediterranean Province. A total of 14 genera, including Brevitheca and Chimerolites, which are restricted to the Caradoc,

217

218

 .   . are found in this interval. Quadrotheca occurs in the Mediterranean Province in the Late Ordovician but is known from the Mid Ordovician in the Baltic Province. Hyoliths were reported from the Late Ordovician of Sweden (Holm 1893), but none of those specimens may be confidently assigned to a genus (Malinky unpubl.). Decipilites, Mediolites, and Railites are the youngest Mediterranean hyolith genera to appear and are confined to the Ashgill (TS.6a–c). No Ordovician genera are known from the Late Ordovician in Laurentia. Species Diversity The endemic nature of hyolith species is even more pronounced than that of the genera, with no species common to both Mediterranean and Baltic provinces or to any other region. As with the hyolith genera, the apparent temporal restriction of species is partly an artifact of taxonomic treatment, as many taxa from outside the Baltic and Mediterranean provinces await reexamination. The possibility definitely exists that even species may have wide geographic distributions, as suggested by the Mid Cambrian Contitheca cor from Sweden, which seems to be present in Korea as Hyolithes kotoi. At present, all known Ordovician species appear to be confined to their respective provinces. The pattern of hyolith species biodiversity (figure 22.2) follows that of the genera in that diversity is lowest in the Tremadocian (six species), followed by an increase to eight at the Tremadoc-Arenig boundary. Diversity increases slightly at the Lower Ordovician– Middle Ordovician boundary to nine species. The Tremadocian species consist of four Baltic and two Mediterranean elements, whereas among the later Early Ordovician species only two come from the Baltic, with all others from the Mediterranean Province. One Laurentian species extends through the entire Early Ordovician. In the Mid Ordovician, two Baltic genera are known from near the Volkhov-Kunda (TS.4a) boundary, with the total increasing to eight in the Kunda. Hyolith biodiversity substantially increases in late Mid Ordovician, with 16 species from the Mediterranean Province and nine from the Baltic. The highest level of diversity is attained in the Upper Ordovician (Caradoc/Viru) with a total of 38 species, followed

by a decrease to half that number in the Ashgill. All Late Ordovician species are from the Mediterranean Province. ■

Cornulitids, Coleoloids, and Sphenothallids (MAW)

Cornulitids, coleoloids, and sphenothallids are extinct tube-dwelling groups sometimes abundant in Ordovician faunas. They are not related to one another, at least not closely, and are treated here together for convenience. The systematic positions of these three groups are poorly understood, and their preserved skeletons are so simple that distinguishing species is difficult, but recent work has shown that they can be important for paleoecologic and paleoenvironmental interpretations, especially in the Ordovician.

Cornulitids The family Cornulitidae Fisher 1962 currently consists of four genera (three in the Ordovician) and about 45 species of small to medium-sized (2–80 mm) calcareous tubes variously ornamented with concentric rings and longitudinal striae. Although they resemble “worm tubes” such as those of serpulids and spirorbids, the vesicular microstructure of cornulitid walls is very different from the laminar microstructure of annelid tubes, leaving their placement in a higher classification uncertain (Fisher 1962). Cornulitids are most commonly found cemented to hard substrates such as shells, carbonate hardgrounds, and rock grounds, but a few occur as unattached single tubes or clusters (Morris and Felton 1993). The earliest cornulitids occur in the Upper Ordovician (Mohawkian), and the latest members of the family are found in the Tennessean of the Lower Carboniferous (Richards 1974). Ordovician cornulitids were almost certainly cosmopolitan in their distribution but thus far have been reported only from North America and Europe. The three Ordovician cornulitid genera are Cornulites Schlotheim 1820. Solitary conoidal tubes with prominent concentric rings and longitudinal striae in adult forms; vesicular walls thick; tubes reach 80 mm in length and 20 mm in diameter at aperture; Upper Ordovician (Mohawkian) to

Tube-Shaped Incertae Sedis Middle Devonian and possibly Lower Carboniferous (Fisher 1962). Conchicolites Nicholson 1872. Conoidal tubes found solitary or in clusters; tubes attached at the narrow tip, which is slightly curved; short, imbricated rings but no longitudinal striae; vesicular walls thin; tubes up to 13 mm long, with diameters up to 3 mm; Upper Ordovician (Mohawkian) to Lower Devonian (Fisher 1962; Richards 1974). Cornulitella Howell 1952. Solitary conoidal tubes attached along one side; rings prominent on all sides except the attached one; no longitudinal striae; thick vesicular walls; tubes up to 13 mm long and 3 mm in diameter; Upper Ordovician to Lower Carboniferous (Fisher 1962).

The paleoecology of Ordovician cornulitids has been extensively treated by Morris and Rollins (1971), Richards (1974), and Morris and Felton (1993). Most cornulitids attached to hard substrates and were undoubtedly filter feeders, but a few lived unattached in soft sediments, leading Richards (1974:515) to consider them as possible “experiments in deposit feeding.” Morris and Felton (1993) showed a convincing symbiosis between the crinoid Glyptocrinus, platyceratid gastropods, and Cornulites in the Cincinnatian of the upper midwestern United States.

Coleoloids The family Coleolidae Fisher 1962 was placed “provisionally” in the phylum Mollusca by Fisher (1962). Coleoloids are elongate cones made of calcium carbonate (calcite, but some may have been aragonitic) ranging from 0.5 to 75 mm long. They usually have a slight curve toward the closed apex. The tubes are unornamented, or have longitudinal striae, or bear oblique ridges. The shell walls are relatively thick and laminated, with some interiors in most. About 20 species have been described in seven coleoloid genera (Fisher 1962), but several of these may belong to other groups of conoidal organisms. The earliest coleoloids appear in the Lower Cambrian, and the last of them are found in Carboniferous rocks. Coleoloids are usually very rare, but they are occasionally found in oriented masses (Fisher 1962; Yochelson 1968). Three coleoloid genera have been reported from Ordovician rocks. One of them, Polylopia, was later

removed from the coleoloids by Yochelson (1968). It is included here for historical reasons and to give it a temporary home for the discussion of Ordovician biodiversity. Paoshanella Yin 1937. Compressed cone, lenticular in cross section; longitudinal striations; up to 60 mm long; Lower Ordovician of China (Fisher 1962). Polylopia Clark 1925. Straight cone with longitudinal ribbing and circular cross section; up to 3 cm long (Fisher 1962); ?Middle Ordovician of North America. Polylopia was originally described as “multilayered”; the name itself is derived from the Greek for “many layers of tree bark” (Clark 1925). Yochelson (1968) demonstrated, though, that this was a misinterpretation of depositionally nested, single-layer cones. Yochelson (1968) made a strong case that Polylopia is a mollusk and, based on more tenuous morphological and paleoecologic evidence, suggested that it might be a hyolith. Salopiella Cobbold 1921. Small (up to 3 mm long) with oblique steplike ridges around the tube on the interior as well as the exterior; Lower Cambrian of England (Cobbold 1921) and “lower part of Middle Ordovician” of Volynia, Ukraine (Hynda 1973:250). Fisher (1962:W134) considered the coleoloid placement of Salopiella “uncertain” because the interior of the shell is not smooth.

Polylopia is the only one of these genera for which we have paleoecologic information. Yochelson (1968) interpreted Polylopia as a benthic organism living in very shallow nearshore marine waters. It may have carried its shell erect, aided by gas inside the closed apex.

Sphenothallids Sphenothallids consist of a single genus (Sphenothallus Hall 1847) with a complicated taxonomic history reviewed most recently by Zhu et al. (2000). Their amended diagnosis describes Sphenothallus as an elongate, slender theca, single or branched, with a subconical holdfast and two longitudinal thickenings extending up the theca; the theca is composed of apatite or organic material, and it is lamellar with the lamellae parallel to the theca surface; the thecae range in length from less than 2 mm to tens of mm. Sphenothallus occurs from the Lower Cambrian to the Permian in North and South America, Europe,

219

220

 .   . and southern China (Zhu et al. 2000). It is found throughout the Ordovician, beginning in the Tremadocian (Choi 1990), with the most common occurrences in the Cincinnatian of North America (Bodenbender et al. 1989; Bolton 1994; Neal and Hannibal 2000). Most commonly sphenothallids are seen in the Ordovician as clusters of holdfasts (Bodenbender et al. 1989; Neal and Hannibal 2000). The systematic placement of Sphenothallus is uncertain. Two hypotheses are current: that they were “worms” of some sort (i.e., Mason and Yochelson 1985), or that they were cnidarians (i.e., Van Iten et al. 1992). The branching, clonal nature of Sphenothallus, along with the similarity of its walls to those of conulariids, favors the cnidarian hypothesis. Sphenothallus was apparently a gregarious, euryoxic, and eurytopic organism comfortable in a variety of marine conditions, from soft muds to shells and hardgrounds (Neal and Hannibal 2000). ■

Byroniids (LEH)

Byroniids (order Byroniida Bischoff 1989) are phosphatic and/or organic tube-shaped fossils that were attached to the substrate by a disk (generally between 0.1 and 0.8 mm in diameter); they have been described from most continents and range in age from the Cambrian to the Permian (Bischoff 1989). Fragmented remains of the isolated phosphatic attachment disks (figure 22.3A) have generally been referred to Phosphannulus Müller, Nogami, and Lenz 1974, but this genus is now considered to be a junior synonym of Byronia Matthew (see Bischoff 1989: 477). In the Early Paleozoic, these attachment disks have generally been referred to a single species, Byronia universalis (Müller, Nogami, and Lenz 1974), which ranges from the Late Cambrian to the Late Devonian. The attachment disks of byroniids are closely similar to those of Sphenothallus Hall, but the internal structure of the tube appears to be different (see Wilson earlier in this chapter).

Ordovician Byroniids Byroniids appear to represent a moderately common constituent in Ordovician shallow-water carbonate deposits from across the world, but to date there are only a very limited number of published

FIGURE 22.3. Byronia universalis Müller et al. (1974). A, B, Folkeslunda Limestone (sample DLK-Fo-5; Holmer 1989: figure 9A), Lasnamägi regional stage, Dalarna, Sweden. C, D, Viivikonna Formation, Kukruse regional stage, Kohtla, Estonia. A, isolated attachment disk, ×35. B, curved tube and attachment disk, ×10.5. C, tubes and attachment disks on dorsal valve of Schizotreta sp., ×2.9. D, isolated tube, ×10.5.

accounts, and there is no basis for analyzing the patterns of their distribution. The first account of an Ordovician byroniid seems to have been by Öpik (1930), who described and illustrated the occurrence of numerous attachment disks (Öpik 1930:31, figure 11, plate 5:2) of what are evidently “Phosphannulustype” byroniids (probably representing Byronia universalis in a wide sense) from the Kukruse beds (lower Upper Ordovician, TS.5a) in Estonia (figure 22.3B, C). Koslowski (1967) described well-preserved organicwalled byroniids from Baltic erratic boulders of Caradoc age occurring in Poland (TS.5; see also Bischoff 1989). Warn (1974) extended the record to North America and described material now referred to byroniids from the Late Ordovician of Ohio. Müller et al. (1974) completed a comprehensive study of the Early Paleozoic record of “Phosphannulus-type” byroniids; they recorded B. universalis from the Late Cambrian–Early Ordovician of Wyoming, Iran, and Sweden. In Sweden there is an almost continuous record of forms referable to B. universalis, which ranges at least from the Tremadocian (TS.1b; Müller

Tube-Shaped Incertae Sedis et al. 1974) through the Volkhov-Kunda (TS.3a–4b; Müller et al. 1974; Eisenack 1978; and Holmer unpubl.) and the Aseri-Kukruse regional stages (TS.4c– 5a; Müller et al. 1974; Holmer unpubl.; figure 22.3A, B herein) to the late Caradoc (TS.5b–c; Holmer 1986, 1987). Much of the Baltic material of byroniids is well preserved and includes both the attachment disk and the long tube (e.g., Holmer 1987: figure 1K, L; figure 22.3A, B herein). Bischoff (1989) described an extensive material of Early Paleozoic Australian byroniids including a single species from the Late Ordovician of New South Wales; he also reviewed the entire record of Early Paleozoic byroniids. In Canada, byroniids have been noted from the Cambrian-Ordovician (Landing et al. 1980) and the Ordovician-Silurian boundary beds (Nowlan et al. 1988) in the Northwest Territories.

Mode of Life and Zoological Affinity The remains of byroniids are generally found as a by-product of etching for conodonts or other acidresistant microfossils (e.g. Bischoff 1989; Müller et al. 1974), and thus any evidence of the host or substrate to which the attachment disk was cemented is lost. The best-known cases of still-attached byroniids are from crinoid stems, where “Phosphannulus-type” byroniids with phosphatic disks and tubes are surrounded and embedded within the stereom of the stem, forming so-called stem galls; the record of this association ranges from the Late Ordovician (Warn 1974) to the Permian (Welch 1976; see also Werle et al. 1984). Müller et al. (1974) described two possible attached disks of B. universalis sitting on an indeterminate phosphatic fragment and a poorly preserved lingulate brachiopod. The association between lingulate brachiopods and attached byroniids is also known from the Kukruse beds (lower Upper Ordovician, TS.5a) in Estonia, where dorsal valves of the discinid brachiopod Schizotreta are covered by numerous disks and tubes of B. universalis (figure 22.3C). However, the byroniids do not appear to have been hostspecific, since identical attached phosphatic disks have also been reported occurring on trilobites and other types of brachiopods from the same level in Estonia by Öpik (1930: figure 11, plate 5:2). The byroniids have been considered to represent a variety of different animal groups but most com-

monly referred to some kind of tube-forming worm (e.g., Müller et al. 1974). However, Bischoff (1989) summarized and reviewed the contrasting views on the zoological affinities of byroniids and concluded that the tube and attachment disks of byroniids are most similar to the attached thecae of the polypoid stage of modern coronate scyphozoans (see also Glaessner 1971, 1984; Bischoff 1978). The Byroniida can be considered as an extinct order of the thecate scyphopolyps (Bischoff 1989). ■

Tentaculitids (HL)

If one accepts the most generally adopted classification of tentaculitids (e.g., Lardeux 1969; Larsson 1979), namely, as the class Tentaculitoidea Lyashenko 1957 consisting of three orders—Tentaculitida Lyashenko 1955, Homoctenida Boucek 1964, and Dacryocorarida Fisher 1962—it may be suggested that no occurrences of indisputable Ordovician tentaculitids have been published yet. These claims rest on two bases: some authors class as tentaculitids certain shells that undoubtedly resemble the genus Tentaculites but whose state of preservation is insufficient to allow this attribution with any certainty; and others, more frequently, assign shells of the cornulitid group (e.g., Cornulites, Conchicolites) to the tentaculitids. Since the insightful contribution of Yochelson (1961, 2000), followed by Fisher (1962), it has been generally agreed that the hyolithids must be distinguished from the tentaculitids. Fisher (1962), for his part, distinguishes between tentaculitids and cornulitids, for which he created a new family, placed in an “uncertain” order and class. Boucek (1964), however, took a different view, proposing a new class, Tentaculita, which combined the orders Tentaculitida, Homoctenida, Dacryoconarida, Coleolida, and Cornulitida. It is true that some cornulitids are strongly reminiscent of tentaculitids, a fact that was first noted a long time ago (e.g., Hall 1847; Barrande 1867). Nevertheless, Boucek’s suggestion has not gained general acceptance (see Lardeux 1969; Larsson 1979). The assumption of the presence of tentaculitids at the beginning of the Early Ordovician rests on the discovery of small conical shells with ringed exteriors, but lacking their apical and oral regions, from the Chepultepec Limestone (early Tremadocian age) of

221

222

 .   . the Shenandoah Valley (Virginia, United States). These shells were called Tentaculites lowndoni by Fisher and Young 1955 (but named “Tentaculites” lowndoni by Fisher in Downie et al. 1967). I have had the opportunity to examine a specimen of this species in the Smithsonian Institution (ref. U.S.N.M. 33403), and my judgment is that its state of preservation is too poor for it to be identified as a tentaculitid. It would appear that it is on the basis of the work of Fisher (1962) that several authors cite the presence of tentaculitids in the Ordovician. However, a closer reading of Fisher’s text reveals that he very clearly allows that the appearance of Tentaculites and Uniconus in the Lower Ordovician is subject to doubt, although this is not represented in the distribution table he provided (Fisher 1962: figure 53). We see this doubt though, on pages 110 and 113, where the Lower Ordovician distributions of the two genera are marked by question marks. The only species he represents as appearing uncontroversially in the Ordovician (p. 111) is “Tentaculites anglicus Salter” from the Caradoc of England, but this may be a cornulitid, as already suggested by Boucek (1964:50). It is true that the “Tentaculites” anglicus shell is nearly as straight as that of true tentaculitids, but its marked longitudinal costulation, a feature of cornulitids, should not be confused with the discreet microcostulation of tentaculitids (Lardeux 1969: plate IX, figure 1). From reports in the literature it may be suggested that the first uncontroversial tentaculitids appeared toward the middle of the Silurian (see Larsson 1979). However, Bergström (1996) has confirmed the presence of tentaculitids in the Upper Ordovician rocks of Ohio. The two illustrated species are relatively common in strata of Richmond age (i.e., ranging through from Waynesville to Whitewater formations). One of the species, Tentaculites richmondensis Miller, exhibits straight to slightly curved, tapering conchs, 25 to 30 mm long and 2 to 3 mm wide, with distinctive transverse rings that gradually increase in size from proximal to distal ends (no bulbs), and the proximal portions of conchs show internal septa. Only the well-preserved specimens show an ornamentation of fine longitudinal striae. The other widespread species,

T. sterlingensis Meek and Worthen, is much smaller and has more densely spaced transverse rings. Tentaculitids also occur in slightly older Cincinnatian rocks, for example, in the Kope Formation of Ohio (S. M. Bergström pers.comm.). These are of Edenian (late TS.5c–early 5d), or late Caradoc, age. In addition, in recent discoveries, as yet unpublished, Kent Larsson (pers. comm.) has identified Caradoc tentaculitids from the Onny Valley of southern Shropshire, England, including specimens of “Tentaculities” anglicus, with all the typical features of the tentaculitid group. They exhibit multilayered, unattached, conical conchs with transverse rings, internal septa, and closed apical ends that are tapered without bulbs. Furthermore, these Ordovician tentaculitids possess a moderately distinct longitudinal costulation, and so this feature is not typical only of cornulitids, and they may also show a slight curvature apically, like many Silurian tentaculitids. Larsson has concluded, based on this Onny fauna, that true tentaculitids were present by mid Caradoc time. Additionally, Kent Larsson found Ashgill tentaculitid assemblages in Sweden and Norway, and some of these specimens have a large size (40 to 60 mm in length). Another occurrence of Ashgill tentaculitids (also currently undocumented) is in thin-bedded limestones (unit 6) of the Tsagaandel Formation at Tsagaandel Hill, west of Bayankhongor, central Mongolia (B. D. Webby pers. comm.; for stratigraphic context, see Minjin 2001). Consequently, the tentaculitids first appeared, like many other Paleozoic groups, during the Caradoc and have a continuing presence in the Ashgill. The earlier Ordovician records of tentaculitids, however, remain unsubstantiated. Affinities of tentaculitids remain uncertain; possibly they represent an independent group long separated from major phyla such as the Annelida and Mollusca.

 Kent Larsson (Lund, Sweden) kindly allowed his new and significant confirmation of occurrences of Late Ordovician tentaculitids in the British Isles and Baltoscandia to be mentioned here.

23

Worms, Wormlike and Sclerite-Bearing Taxa Olle Hints, Mats Eriksson, Anette E. S. Högström, Petr Kraft, and Oliver Lehnert

T

he biodiversity of four wormlike groups is outlined in this chapter. The first is a larger group, the jaw-bearing polychaetes, which are represented by scolecodonts, the organic-walled elements of their jaw apparatuses. The others are smaller, more problematic groups. The machaeridians are most commonly preserved as isolated calcitic sclerite, derived from the dorsal exoskeletal armor of an unknown, bilaterally symmetrical wormlike organism. The palaeoscolecideans and chaetognaths (arrow worms) are phosphatic, identified mainly by their microfossil remains and a few by more complete body fossils. Characteristic microfossil elements of palaeoscolecidans are platelike tubercles, and the chaetognaths exhibit spinelike grasping elements that suggest the chaetognaths may have close links to protoconodonts. ■

Jawed Polychaetes (Scolecodonts) (OH, ME)

Scolecodonts, the jaws of polychaete worms (Annelida, Polychaeta), are common microfossils in Ordovician sedimentary rocks. They are composed of yellow to black organic material allowing acid preparation techniques to be used for their extraction. The first report on Ordovician scolecodonts is by Hinde (1879), who described specimens from Canada. Since then, Ordovician scolecodonts have been recovered from many regions of the world, for

example, Australia (Furey-Greig 1999), Kazakhstan (Klenina 1989), China (Gao 1980), South America (Ottone and Holfeltz 1992), North America (e.g., Stauffer 1933; Eller 1945 1969; Bergman 1998; Eriksson and Bergman 1998, 2001), and Europe (e.g., Kielan-Jaworowska 1966; Szaniawski 1970; Hints 1998). However, except for the last two regions, the existing data are far too limited to allow detailed diversity analyses. Hence, this chapter is focused on the Baltic region and North America. Preliminary data on interregional distribution of jaw-bearing polychaetes indicate that there were close links between the faunas of different continents during the Ordovician (Hints et al. 2000; Eriksson and Bergman 2001). Most genera were common to Laurentia and Baltica, but the majority of the few species recovered from other regions apparently also belong to these widespread genera. There are, however, a number of Baltic genera that have not yet been identified in Laurentian deposits and vice versa. At the species level, on the other hand, relatively few forms are present in more than one continent. Thus, currently available species-level diversity data characterize a particular region, or a continent, but cannot yet be used for the entire world. Consequently, the global diversity of jawed polychaetes can be assessed in a meaningful way only at the genus level, and even then, the preliminary nature of the data must be stressed. 223

15

30 JAWED POLYCHAETE GENERA (Range-through data)

Normalized diversity (dnorm) Genera per million years (di) Rate of origination (oi) Rate of extinctions (ei)

10

5

20

10

Normalized diversity and Genera per million years

   . Turnover (rates per million years)

224

0

0 1a

1b

1c

1d

2a

2b

Tremadoc Ibexian C

2c

3a

Arenig

TREMADOCIAN

489 Ma

3b

4a

4b

4c

Llanvirn Whiterockian

5a

5b

5c

Caradoc Mohawkian

5d

6a

6b 6c

Ashgill Cincinnatian

DARRIWILIAN

LOWER ORDOVICIAN

MIDDLE ORDOVICIAN

472

460.5

UPPER ORDOVICIAN

S

443

FIGURE 23.1. Genus-level diversity pattern of Ordovician jawed polychaetes. Stratigraphic scale is given in accordance with Webby et al. (chapter 2).

A multielement taxonomy has been applied here for analyzing the distribution and diversity patterns of jawed polychaetes. Taxa recognized using the outdated single-element taxonomy are of little use without careful revisions of old literature and the corresponding type collections (Eriksson and Bergman 1998). The taxonomic richness was calculated using normalized diversity and taxa per million years. The normalized diversity (number of taxa ranging through the time unit, plus half the number of taxa appearing or disappearing in it, plus half the number of those confined to the unit) differs from total generic diversity only by producing slightly lower values and a smoother curve (figure 23.1). In addition, origination and extinction rates per million years were calculated. In all cases, range-through data rather than actually recorded data were used. Approximately 50 Ordovician genera, including those undergoing taxonomic revision and those that lack formally proposed generic names, are currently known. The number of families is debatable, but some 15–20 families were probably present in the Ordovician. The total number of species is even more difficult to estimate. In the Baltic region, 150 multielement species are known. The number of species seems to be slightly lower in North America, and some 70–100 species have thus far been recovered, based mainly on material from the midcontinent region. These numbers will most likely increase with ongoing sampling and investigations combined with revisions of the pioneer scolecodont literature. The oldest known scolecodonts have been recovered from the uppermost Cambrian (S. H. Williams

pers. comm.) in Newfoundland. The same area has yielded some of the oldest known Ordovician scolecodonts (Underhay and Williams 1995; Williams et al. 1999), but Early Ordovician specimens have also been found in Estonia by O. Hints (unpubl. data) and in China (R. Brocke pers. comm.). The collections from TS.1–2 (see time slices in chapter 2) have not yet been thoroughly studied, but it seems that the diversity of jawed polychaetes is very low through this interval. The limited number of jaw apparatuses recovered or reconstructed allows at least three distinct genera to be distinguished. It is currently premature to evaluate the higher-level taxonomy of early polychaetes, although it seems that some of these forms are related to xanioprionids and conjungaspids and others are very similar to Lunoprionella, a common genus in younger Ordovician strata. More advanced forms, such as those with labidognath and prionognath type apparatuses (e.g., Kielan-Jaworowska 1966), have not been recorded in this interval. In TS.3a, members of Oenonites and Mochtyella are first recorded (Underhay and Williams 1995). These genera, particularly the former, commonly dominate the assemblages in younger Ordovician strata as well as in the Silurian (Eriksson 1997; Hints 2000). At least by this time (TS.3a) the evolution of jaw-bearing polychaetes had passed some important milestones, and the main polychaete lineages had become differentiated. The diversity remains rather low until TS.4, when a major increase in species diversity, as well as in abundance, is recorded and the number of genera increased very rapidly. Moreover, this interval marks

Worms, Wormlike and Sclerite-Bearing Taxa the first appearance of most of the genera that became common in younger strata. A major component of the rapid increase in normalized diversity occurs during the TS.4c–5a interval (figure 23.1). Since the genera appearing in TS.4c and 5a commonly range through the remaining part of the Ordovician, the diversity curve changes only very little in later time slices, displaying a slight increase until TS.5d–6a and a subsequent decrease in TS.6b–c. It is difficult to ascertain whether minor fluctuations are meaningful. Since most genera occurring in the uppermost Ordovician range into the Silurian, there is also no distinct drop in diversity at the Ordovician-Silurian boundary, which is marked by a significant diversity drop in many other fossil groups. Genera per million years gives a pattern slightly different from the normalized diversity curve (figure 23.1). It shows much smoother increasing trend with some slightly steeper parts in TS.4c and 5b. However, as most genera occur in several time slices, the taxonomic richness in shorter time slices, such as TS.6b and 6c, becomes somewhat overestimated using this measure. The turnover estimates are in good accordance with the normalized diversity curve. There is a small peak in origination rate in TS.3a, but the major peak is confined to TS.4c. After that the origination rate remains very stable until a decrease in TS.6b and 6c (see figure 23.1). The extinction rate, in contrast, has no clearly defined peaks. That is, jawed polychaetes appear to have been relatively unaffected by some of the events causing severe extinctions in many other fossil groups. Composite species-level data, such as those provided by Hints (2000) for the Baltic region, may display higher complexity, especially in turnover rates. However, as is the case in global genus-level data, the species-level taxonomic richness in the Baltic region remained relatively stable after the main diversity rise (Hints 2000: figure 2). There is some question as to how well the revealed pattern corresponds to the actual diversification of the group and whether conclusions can be drawn based on such a patchy data set. It is indeed very likely that the number of genera will increase and that the ranges of many taxa will be extended with results from ongoing investigations. Therefore, the present diversity estimates should be viewed as preliminary rather than comprehensive (i.e., the diversity was at

least not lower than here indicated). The record is especially incomplete for the Lower and lower Middle Ordovician (TS.1–3). Further data from this interval will be invaluable, and they will likely change the current picture to some extent. The present knowledge about the diversity patterns of Ordovician jaw-bearing polychaetes can be summarized as follows: 1. Jawed polychaetes originated in pre-Ordovician time. 2. Several advanced taxa appeared in TS.3, but the most significant increase in diversity and abundance seems to have occurred in the Middle Ordovician (TS.4), after which genus-level diversity remained relatively stable without distinct peaks in origination or extinction rates. 3. Most of the Ordovician genera are apparently long-ranging, and many of them extend into the Silurian. 4. An intercontinental faunal exchange is indicated, and most families and many genera were common to Laurentia and the Baltic region. However, there are taxonomic differences in the faunas between these paleocontinents, especially at the species level. 5. The abundance and great diversity of scolecodonts in the Ordovician warrant further investigation of this group. ■

Sclerite-Bearing Machaeridians (AESH)

Sclerite-bearing taxa of varying types abound in Lower Paleozoic rocks from many parts of the world. Several have a very restricted stratigraphic range, and a number of better-known groups, such as the tommotiids and halkieriids, occur only in the Cambrian. The problematic machaeridians, however, have a substantial stratigraphic distribution (Lower Ordovician to Middle Permian). Additionally, their disarticulated remains are being recognized and reported in increasing numbers (Ordovician examples include Dzik 1994b; Ebbestad and Högström 1999; Högström and Droser 2001). All scleritome-forming taxa suffer essentially similar preservational problems: the sclerites become widely dispersed after the animal dies, and isolated sclerites are thus by far the most common type of preservation. Hence, the distribution of isolated sclerites is the main indicator of their spatial and temporal

225

226

   .

FIGURE 23.2. A selection of different Ordovician machaeridian types. A, Middle Ordovician Lepidocoleus ulrichi Withers 1926, anterior end of a partial scleritome, from the Trenton Group, Prosser Limestone, Cannon Falls, Minnesota, United States, ×5.6. B, Plumulites sp. from the upper Ashgill of the Taimyr Peninsula, Arctic Russia, ×8.0. C, Plumulites sp. from the Ashgill Boda Limestone of central Sweden, ×2.1. D, E, Lepidocoleus suecicus Moberg 1914, from the Ashgill of Bohemia, latex casts, both ×6.3. F, G, minute plumulitid-type sclerites with extreme marginal spines from the Ashgill, Fjäcka Shale, Koängen boring, Scania, Sweden; F, ×16.7; G, ×10.5. H, I, plumulitid-type machaeridians from the upper Ibexian (Arenig) Al Rose Formation of the Basin and Range, Inyo Mountains, California, United States, both ×4.5.

distribution as well as of their faunal importance and environmental preferences. However, complete scleritomes are necessary for accurate reconstructions and for understanding the morphology and function of the scleritome. In general a machaeridian scleritome consists of two or four longitudinal series of posteriorly overlapping and imbricating calcitic sclerites (see figure 23.2A). These are arranged in transverse segments of varying numbers, from 14 to well over 60 (Högström and Taylor 2001). The three major families (Lepidocoleidae, Plumulitidae, and Turrilepadidae), grouped in the two orders (Lepidocoleomorpha and Turrilepadomorpha), are primarily distinguished on the basis of scleritome arrangement ( Jell 1979). Machaeridians have a wide environmental distribution that ranges from deep offshore shelves to shallow epicontinental seas and includes apparent epibenthic as well as semi-infaunal modes of life (Högström 2000). An example of the former may be the extremely spiny and minute sclerites of an undescribed plumulitidlike machaeridian from the Upper Ordovician Fjäcka Shale of Sweden (figure 23.2F, G).

When the organism is reconstructed, the resulting “snowshoe effect” conferred by these spines can be interpreted as a possible adaptation to life on unstable soft substrates. Terrace-shaped rugae reminiscent of burrowing sculpture found in some machaeridians suggest a possible semi-infaunal lifestyle (for example, see figure 23.2A, D–E). In addition, a large collection (∼100 specimens) of well-articulated lepidocoleid machaeridians from the Lower Devonian (Lochkovian) of Oklahoma are preserved in a manner indicating rapid burial and possible in situ preservation within fine-grained muds. More unusual is the occurrence, from carbonate mud mounds of the Upper Ordovician Boda Limestone of Sweden, of rare sclerites of a very large plumulitid machaeridian (figure 23.3C), in an environment otherwise seemingly devoid of machaeridians.

Ordovician Taxa Despite the increasing number of reported machaeridians, their record is still “spotty” and with tax-

Worms, Wormlike and Sclerite-Bearing Taxa onomic uncertainties largely caused by preservation and relatively few specimens. It is, however, valuable to produce a general compilation of machaeridian distribution through the Ordovician, based on work published so far as well as on personal observations and communications. Tremadocian The oldest confirmed machaeridian sclerites are of plumulitid type from the Tremadocian Dumugol Formation (Dongjeom area) of South Korea (Kobayashi and Hamada 1976; Choi and Kim 1989), where Plumulites gumunsoensis Choi and Kim 1989 was quoted as a fairly common but overlooked component of the fauna in the Dumugol Formation. From the same formation Plumulites primus Kobayashi 1934 is known (Kobayashi and Hamada 1976). The Bjørkåsholmen Formation of Norway may contain the oldest Baltic machaeridians (B. Funke pers. comm.). Arenig The early to mid Arenig shows signs of the first larger radiation of machaeridians with several new occurrences of plumulitid-type machaeridians, for example, in the Al Rose Formation of the Inyo Mountains of the Basin and Range, California (figure 23.2H, I) (pers. observ.). Barrande (1872) described the earliest machaeridians from Bohemia (four species of Plumulites), but there is a slight degree of uncertainty in their stratigraphic placement, and some may be of late Tremadocian age. Toward the upper Arenig, machaeridians are known also from northeastern North America (Clark 1924), China (Kobayashi and Hamada 1976), and the Holy Cross Mountains of Poland (Dzik 1994b). In addition, as yet undescribed material from the St. Petersburg district consists of what may be the earliest lepidocoleidtype machaeridians (pers. observ.), belonging to the Baltoscandian middle and upper Volkhov stage (Egerquist 1999). Llanvirn The trend from the Arenig continues into the Llanvirn (mid to late Darriwilian) with an increase in the number of taxa and their distribution. A species of Plumulites occurs in abundance in the Kanosh For-

mation of the Ibex area in the Great Basin of Utah and Nevada. Machaeridians also show a continuous distribution through the Holy Cross Mountains sequence (Dzik 1994b). The first machaeridians from Morocco (Chauvel 1967) appear in the Llanvirn and show a continued Gondwanan distribution, and the interval also marks the continued appearance of South Korean and Chinese material (Kobayashi and Hamada 1976). Eastern North America continues to show the same increasing trend with additional plumulitid machaeridians as well as the first lepidocoleids (Hall and Whitfield 1875; Withers 1926). One of the first reports with certain Baltic affinity is from the Kukruse stage in Estonia (Withers 1921), and they also continue to occur in Bohemia. Caradoc The Late Ordovician exhibits taxa persisting from the Llanvirn, for example, from the Holy Cross Mountains (Dzik 1994b), but also from South Korea (Kobayashi and Hamada 1976), North America (Withers 1926), and Morocco (Chauvel 1967). New taxa are recorded from among other areas—Sweden, Bohemia, North America, and Scotland (Barrande 1872; Moberg 1914; Withers 1926). Overall the Late Ordovician is a time when species diversity increases and machaeridians appear to thrive in many settings. Ashgill The uppermost Ordovician contains faunas with diverse and numerous machaeridian remains (Dzik 1994b; Ebbestad and Högström 1999). In Europe several taxa from the Holy Cross Mountains continue into the Ashgill (Dzik 1994b), as well as Bohemian and Baltic taxa (Barrande 1872; Withers 1926). The Baltic fauna, especially the Fjäcka Shale, shows a diverse and conspicuous machaeridian component, where they may be the numerically dominant taxon in some beds. In addition to plumulitids and lepidocoleid machaeridians there are examples of new minute forms with extreme marginal spines mentioned earlier (figure 23.2F, G). The North American fauna continues to exhibit Lepidocoleus jamesi (Hall and Whitfield 1875) and Lepidocoleus strictus Withers 1926, as well as components that at this point appear relatively similar to Baltic types. The occurrence in China of two subspecies of a Baltic species indicates a

227

228

   . possible faunal connection between these areas during the Ashgill (Wu 1990). Additional Malaysian and South Korean taxa occur in the Upper Ordovician as well, giving this region a continuous record from the Tremadocian (Kobayashi and Hamada 1976). Moroccan machaeridians are seemingly diverse and abundant (Chauvel 1967) but in urgent need of further studies. In Russia Plumulites sp. occurs just below the Ordovician-Silurian boundary layers of the Taimyr Peninsula (figure 23.2B) (pers. observ.).

Discussion The general impression of machaeridian diversification and distribution through the Ordovician and onward is one of continuing success. Although the group is not one of the most common, its substantial stratigraphic distribution and environmental and geographic spread suggest very successful organisms. Uncertainties and problems do remain with the machaeridian record, owing to both preservation and lack of study. However, an increased focus on certain areas has shown that machaeridians very likely were a more conspicuous component of Paleozoic faunas than previously recognized. This is well exemplified, for example, by the recently discovered machaeridians in the Basin and Range of California and the Great Basin of Utah and Nevada. Work in progress there has revealed large numbers of sclerites from the previously mentioned Al Rose (late Ibexian/Arenig) and Kanosh (mid to late Whiterockian/mid to late Darriwilian) formations and additional (but rarer) isolated machaeridian sclerites from the Nine Mile (Whiterock Canyon, Nevada: late Ibexian/Arenig) and Wahwah (Ibex area, Utah: latest Ibexian to early Whiterockian/Arenig) formations (pers. observs.). Yet another example where machaeridian remains are turning up in large numbers and occasionally constitute one of the most numerous of the taxa is the Ashgill Fjäcka Shale of central Sweden (Ebbestad and Högström 1999), even though machaeridians have been known from this level since the beginning of the 1900s (Moberg 1914). Lepidocoleus suecicus Moberg 1914, common in the Fjäcka Shale, is also found at similar stratigraphic levels in Bohemia (figure 23.2D, E), indicating a possible closer relationship to Baltica than was previously thought. Compared with the diversity in both the Al Rose and Kanosh forma-

tions, that of the Fjäcka Shale is much higher, with a few numerous forms and a number of rarer ones. This also seems to be the general trend, with a substantially higher diversity toward the Late Ordovician. In addition, faunas in North America and Europe grow increasingly similar toward the end of the Ordovician as a result of the closure of the Iapetus Ocean. The apparent conservative morphology of machaeridian scleritomes through time is quite striking. This is especially evident in plumulitid machaeridians, in which articulated Ordovician specimens differ little from later forms, for example, from the Devonian ( Jell 1979; Rudkin 2001). This is at least partly reflected by a low higher-level diversity, in which most differences between forms lie in the sclerites themselves and not in the overall organization of the scleritome. The number of new finds from various parts of the world gives hope that the poor resolution of the machaeridian record will improve significantly in the near future. In addition, more studies of continuous sections such as the one by Dzik (1994b) from the Holy Cross Mountains will help to clarify the picture of the distribution and turnover rates of machaeridians as well as their role in different communities. ■

Palaeoscolecidans and Chaetognaths (PK, OL)

Palaeoscolecidans and chaetognaths (arrow worms) belong to separate phyla. Soft-body preservation is extremely rare in both groups. This fact strongly biases our picture of the diversity. However, isolated phosphatic microelements—present in each of these groups—provide a basis for studies of their biostratigraphic and paleogeographic distribution.

Palaeoscolecidans Palaeoscolecidans have been assigned to a number of different phyla. Most authors consider them as annelids (e.g., Robison 1969; Glaessner 1979). Recently, new opinions have emerged and led to a proposal that their systematic position is as an extinct group of nematomorphs (Hou and Bergström 1994) or more probably priapulids (e.g., Conway Morris 1997). Their body is composed of tuberculate annuli with plates of different shape and distribution.

Preservation of body fossils is due to the phosphatized cuticle (probably secondary). The phosphatic tubercles (plates) are the heaviest mineralized parts, and as a consequence of postmortem decay they are often found isolated (see, e.g., Hinz et al. 1990). These microfossils have been described as Hadimopanella Gedik (= Lenargyrion Bengtson), Kaimenella Märss, and Milaculum Müller, while some sets of integrated microelements are referred to Utahphospha Müller and Miller. The microelements have an independent parataxonomic classification. Naturally, attempts to correlate both modes of preservation have been published. For example, Müller and Hinz-Schallreuter (1993) assigned several fragments of body surface to Hadimopanella and Milaculum because of the typical morphology of tubercles (plates). On the other hand, van den Boogaard (1989) described isolated tubercles (plates) as ?Palaeoscolex, based on similarities to material from Bohemia (figure 23.3, small shaded area of “body fossilized” column of Palaeoscolecida). Palaeoscolecidans range from Lower Cambrian to Silurian (Ludlow). In the Ordovician, body fossils have been described from the lower Tremadocian of Wales (Palaeoscolex Whittard 1953; see also Conway Morris 1997), lower Arenig to the lowermost Caradoc of Bohemia (?Palaeoscolex, Plasmuscolex, and Gamascolex; Kraft and Mergl 1989), and the lowermost Cincinnatian of Kentucky and Ohio (“Protoscolex”; for overview and comment, see Conway Morris 1977; Conway Morris et al. 1982). Isolated phosphatic tubercles (plates) have been recorded from Ordovician rocks in different parts of Laurentia and Baltica. There is also considerable potential for information in many conodont collections (e.g., at different universities and in the Canadian and U.S. Geological Surveys), but this material has not yet been studied in detail. Palaeoscolecidans had a wide paleogeographic distribution from tropical warm-water environments (Laurentia) to temperate-water areas (periGondwana). Although the group ranges throughout the whole Ordovician, occurrences reach a maximum in the Lower and Middle Ordovician (see figure 23.3). The diversity of palaeoscolecidans seems to decrease from Cambrian through the Ordovician until they become extinct in the Silurian. This can be illustrated by the number of described genera: their record in-

N. Amer

Worms, Wormlike and Sclerite-Bearing Taxa

3

1

1

1

1

1

1

3

3

3

3

3

FIGURE 23.3. Stratigraphic ranges of palaeoscolecidans and chaetognaths. Bodily preserved taxa are in the white columns, microelements are shaded. Points mark only important appearance data. Abbreviations: Lower Cambrian (–C1), Upper Cambrian (–C3 ), Lower Silurian (S1), Upper Siluarian (S3 ), Devonian (D).

cludes 15 genera of bodily preserved worms and 4 genera of tubercles (plates) from the Cambrian, 4 (or maybe 5) and 4 genera of worms and plates, respectively, from the Ordovician, and only 1 and 2 genera, respectively, from the Silurian.

Chaetognaths This phylum probably appeared in the Cambrian (e.g., Szaniawski 1982). Today there are some 250 species in this phylum, but soft-body preservation is extraordinarily rare throughout the fossil record. There

229

230

   . are reports from the Lower Cambrian (Protosagitta Hu, in Chen et al. 2002), the Lower to Middle Ordovician, and the upper Pennsylvanian (Schram 1973). The Ordovician material includes three specimens referred to Titerina rokycanensis Kraft and Mergl 1989 (see also Kraft et al. 1999). Globular microfossils, interpreted as eggs of Titerina or a related genus, are known from the Arenig and lower Llanvirn of Argentina (Heuse et al. 1996). This species is unusual in the possession of a prominent pair of grasping spines. Almost identical isolated apparatuses have been documented from the Upper Ordovician of the Canadian Arctic (McCracken and Nowlan 1989). They may belong to Titerina or a closely related chaetognath genus. Despite the paucity of this record, it has been repeatedly advocated that protoconodonts represent the grasping spines of chaetognaths, and we consider they may represent an extinct order of this phylum. Grasping spines in Cambrian and Ordovician protoconodonts are comparable in shape and size to those in modern chaetognath apparatuses and have a similar outline in cross section. In addition, sets of protoconodonts preserved as complete apparatuses have been recorded from the Upper Cambrian, and in particular the affinity between the apparatus of Phakelodus tenuis (Müller) and that of the recent pelagic chaetognath Pseudosagitta maxima (Conant) has been discussed by Szaniawski (1987). In comparison, chaetognath teeth are very small, and in this sense only Szaniawski’s (1996) report of small teeth attached to an Upper Cambrian Phakelo-

dus elongatus (Zhang) cluster shows evidence for the systematic position of protoconodonts within this group. In addition, clusters of the thin-walled elements of Coelocerodontus Ethington, with deep basal cavities, may be attributed to protoconodonts. Their clusters have an apparatus architecture similar to that of Phakelodus Miller (cf. Szaniawski 1998), and the elements are also similar to grasping spines of Titerina. Coelocerodontus is well known from the Upper Cambrian through Upper Ordovician; the youngest reports are from the Devonian (e.g., Telford 1975). Protoconodonts are widespread, diverse, and most common in the Upper Cambrian, but their number decreases at the end of the Cambrian and during the earliest Ordovician. Body fossils and protoconodonts testify to a cosmopolitan distribution of chaetognaths during the Cambro-Ordovician. Their overall diversity during the Ordovician is difficult to estimate. Additionally, the puzzle of isolated protoconodont elements and clusters and their long stratigraphic ranges indicate that chaetognath diversity was probably low in the Ordovician.

 Anette Högström acknowledges support from the Swedish Research Council (Vetenskapsrådet), the National Science Foundation through Mary Droser and the European Community Marie Curie Foundation. Simon Conway Morris and Barry Webby are thanked for their helpful and valuable comments.

24

Trilobites Jonathan M. Adrain, Gregory D. Edgecombe, Richard A. Fortey, Øyvind Hammer, John R. Laurie, Timothy McCormick, Alan W. Owen, Beatriz G. Waisfeld, Barry D. Webby, Stephen R. Westrop, and Zhou Zhi-yi

T

he taxonomic diversity history of Ordovician trilobites has been explored on a broad global scale by Adrain et al. (1998), who provided an estimate based on a fourfold division of Ordovician time. Adrain and Westrop (2000) subsequently published a trilobite diversity curve based on nine Ordovician and five Early Silurian sampling intervals, and the pattern of trilobite alpha (within-habitat) diversity during the Ordovician has been documented by Westrop and Adrain (1998) and Adrain et al. (2000). The nature of the Ordovician radiation of trilobites is further characterized herein, through a new global analysis at finer resolution and by documenting geographic and environmental patterns in the data. Regional diversity curves are presented and discussed for Australasia, South America, Avalonia, Baltica, and South China, reflecting a spectrum of tectonic and paleogeographic settings through the Ordovician. Finally, the development of trilobite biofacies through the period is assessed in the context of the global and regional patterns of biodiversity change.

■

Global Patterns (JMA, SRW, RAF) The Data Set

Data on the taxonomy and temporal and geographic distribution of Ordovician and Silurian trilobites have been compiled by J. M. Adrain beginning in 1997. In their assessment of post-Cambrian trilo-

bite diversity and evolutionary faunas, Adrain et al. (1998) used a database of 1,241 recorded genera, of which they accepted 945 as meaningful taxa—842 are Ordovician genera. At that stage, it was possible to present data only at a relatively coarse series-level resolution. In order to achieve further insight, the resolution of the data set needed to be increased. Effectively, this required that (1) a workable set of intervals be developed that could be applied with a minimum of uncertainty to data from all parts of the world and (2) the data set be extended to species level in order to document genus ranges and geographic occurrence accurately through time. The sampling intervals chosen (see Adrain and Westrop 2000: note 30) were a necessary compromise between the high precision available from some paleocontinents (e.g., Laurentia, Baltica, Australasia, Avalonia, and parts of Gondwana) and the sometimes very coarse resolution in others (e.g., Siberia, the central Asian terranes, and much of South America). One alternative was to adopt a highly resolved scheme and then deal with less-resolved data according to an error distribution formula (the procedure used by Sepkoski 1986; see also Sepkoski and Koch 1996 as a result of dependence on secondary data sources). Although such a system is reasonable for a combined global diversity estimate, it does not allow effective comparison of geographic regions by time interval. We therefore 231

        .           . adopted a system in which data had a direct empirical assignment to a sampling bin, despite the limitations imposed by poorly sampled regions. The result is a ninefold division of Ordovician time, based on major correlative biohorizons identified by Webby (1995, 1998), corresponding approximately to the level of stage or subseries. In terms of the 19 time slices (TS ) adopted for this book, our intervals are less well resolved but at least directly match particular time-slice boundaries. Relationship of the intervals is as follows: O1 = (TS.1a, 1b); O2 = (TS.1c, 1d); O3 = (TS.2a, 2b, 2c); O4 = (TS.3a, 3b); O5 = (TS.4a, 4b, 4c); O6 = (TS.5a); O7 = (TS.5b, 5c); O8 = (TS.5d); O9 = (TS.6a, 6b, 6c). Compilation to the species level is now well advanced. Earliest and latest occurrences have been documented, and there is now considerable confidence in the genus ranges. Overall, the species compilation is sufficiently complete for preliminary analysis, and species data are used to document the latitudinal distribution of families later in this chapter. By 2000, the number of genera recorded had risen to 1383, with 994 accepted. Because about two-thirds of the added names were synonyms or dubia taken from more obscure primary literature, the effective database had increased in size by only 5.7 percent. This is the database analyzed in the present work. A small number of accepted new genera of Ordovician trilobites that were published in 2000 and 2001 are excluded from the present work in favor of concordance with the database used by Adrain and Westrop (2000). Changes in the data set since 1998 include some taxonomic reassignment. Bathyurids and bathyurellids were united in a family Bathyuridae in 1998 but are considered separate families for purposes of this analysis. A family Panderiidae was recognized in 1998, but the group is considered a subfamily of Bathyurellidae in the present data set. The current global trilobite diversity curve based on these data is shown in figure 24.1. A listing of taxa, synonyms, and stratigraphic ranges of the genus data set is available on request from J. M. Adrain.

A New Global Analysis The main conclusion of Adrain et al. (1998) was that a significant portion of post-Cambrian trilobites experienced rapid diversification during the Ordovi-

300 Whiterock Ibex I Ibex II Total

250

Number of Genera

232

200

150

100

50

0 O1 O2 O3 O4 O5 O6 O7 O8 O9 S1 S2 S3 S4 S5

Sample Interval

FIGURE 24.1. Ordovician and Early Silurian trilobite genus diversity. Whiterock, Ibex I, and Ibex II are evolutionary faunas defined by hierarchical cluster analysis (see figure 24.2). Sample intervals are as defined by Adrain and Westrop (2000:112, note 30). It should be noted that sample intervals O1–O3, O4–O5, and O6–O9 are equivalent to the tripartite Ordovician division, respectively, Lower (or Early), Middle (or Mid), and Upper (or Late) Ordovician, used in this volume. Intervals O1 and O5 also correlate with global Stages Tremadocian and Darriwilian, respectively.

cian Radiation in a fashion nearly identical to that of Sepkoski’s (1981a) Paleozoic Evolutionary Fauna. The terms “Ibex Fauna” and “Whiterock Fauna” were introduced. The former comprised a cohort of families that peaked early, declined, and were eradicated before or during the end Ordovician mass extinction; the latter had low Early Ordovician diversity, diversified during the Ordovician Radiation, and contained all families that survived the end Ordovician mass extinction (a group termed the “Silurian Fauna”). Even though the fourfold division used by Adrain et al. (1998) clearly documented major unrecognized features of the trilobite record—Mid Ordovician diversification and its link to end Ordovician survival—a more stratigraphically resolved analysis of the current data set is desirable to test the cohesiveness of the new evolutionary faunas (essentially to see if they break down into discrete components) and better refine the timing of diversification events. The analysis was carried out using the same protocols. Families were grouped according to similarity in their genus diversity through the Ordovician sampling intervals. Silurian data were not used in the analysis. At this increased level of resolution, the data

Trilobites became subject to edge effects. In particular, there were eight families with single occurrences (in some cases of a single specimen) in the earliest part of O1. These taxa essentially have no Ordovician history but, when admitted to the analysis, cluster together with 100 percent similarity and exert undue influence on the pattern of similarity of the remaining (interesting) taxa that have a genuine Ordovician history. These “singleton” taxa (“Dokimokephalidae,” Idahoiidae, Lichakephalidae, Nepeidae, Norwoodiidae, Papyriaspididae, Plethopeltidae, and Solenopleuridae), restricted in occurrence to O1, were therefore excluded from the analysis. The results of the cluster analysis are shown in figure 24.2. The main features are as follows: 1. The Ibex Fauna includes two distinct clusters with different diversity trajectories (see figure 24.1) during the O1–O3 intervals (lumped together as the Ibexian by Adrain et al. 1998). One group, termed “Ibex Fauna I,” had high O1 diversity but steadily declined afterward. The Olenidae and Ceratopygidae are typical of this fauna. A second group, termed “Ibex Fauna II,” had very low O1 diversity but radiated rapidly during O2, peaked during O3, and declined after O5. The Asaphidae and Bathyuridae are typical families of this fauna. 2. The Whiterock Fauna composition remains almost exactly as described in 1998, though there is a small amount of changed membership. The lowdiversity families Dionididae and Bohemillidae, which had clustered with the Whiterock Fauna in the fourinterval analysis, now move to Ibex Fauna II. Isocolidae, which had clustered with the Ibex Fauna, and Harpetidae, the sole unclustered family, both now move to the Whiterock Fauna. Twenty-two of 24 families are common to the 1998 and present versions of the Whiterock Fauna. 3. With increased stratigraphic resolution, it is evident that some Whiterock Fauna families began to radiate during the O3 interval (e.g., Trinucleidae, Raphiophoridae, Cyclopygidae), whereas others began to radiate during the early Whiterockian O4 (e.g., Illaenidae, Encrinuridae, Odontopleuridae). There is a strong geographic component to this distinction, which is discussed later in this chapter. 4. The two most surprising results in the analysis of Adrain et al. (1998) are confirmed. First, the

Whiterock Fauna—the majority of post-Cambrian trilobites—experienced an Ordovician radiation (figure 24.1) much like that of the Paleozoic Evolutionary Fauna. Second, Ordovician diversity history remains an extremely accurate predictor of end Ordovician fate. All families of both Ibex Fauna I and Ibex Fauna II became extinct either before or during the end Ordovician mass extinction, whereas 18 of 26 Whiterock Fauna families survived into the Silurian. There is a strong geographic component to this Whiterock Fauna survivorship, also discussed later herein.

Trilobite Radiation by Realm Modern work on global Ordovician trilobite biogeography (using cladistic methods to search for common historical signals as opposed to phenetic methods to compare taxonomic lists) has not been attempted and is sorely needed. The best estimate of global pattern around the time of the radiation remains Whittington and Hughes’s (1972) classic quantitative study of taxic distribution, which used multidimensional scaling in the first comprehensive attempt to define major biogeographic areas in the Early Ordovician. Their analysis has largely been supported by subsequent work (Ross 1975; Fortey and Morris 1982; Cocks and Fortey 1982, 1988; Neuman 1984; Fortey and Cocks 1986, 1992; Fortey et al. 1989; Fortey and Mellish 1992), with the main additions being attention to the concept of biofacies and accounting for the effects of levels of endemicity varying with environment. Current concepts of broad area relationships during the Early Ordovician have been summarized by Cocks and Fortey (1990). The level of precision in well-studied areas such as eastern Avalonia and western Baltica is now high, though data remain sparse in many other parts of the world. Nevertheless, it is clear that during the time of the radiation, trilobites occupied at least four distinct biogeographic realms (figure 24.3). An equatorial Bathyurid Realm (Bathyurid Province of Whittington and Hughes 1972) includes Laurentia and Siberia/Kolyma and possibly parts of Kazakhstan (as far as is known; see Apollonov 1975 for summary) and North China (Zhou and Fortey 1986). Data adequate for quantitative analysis have been compiled only for Laurentia. The most characteristic preradiation faunal elements are the endemic bathyurids

233

Similarity

Bavarillidae Missisquoiidae Ceratopygidae Eulomidae Hystricuridae Alasataspididae Olenidae Pilekiidae Ptychaspididae Hungaiidae Bathycheilidae Shumardiidae Entomaspididae Leiostegiidae Agnostoidea Pharostomatidae Taihungshaniidae Remopleurididae Dionididae Nileidae Bohemillidae Raymondinidae Bathyurellidae Bathyuridae Diaphanomet. Asaphidae Prosopiscidae Pliomeridae Telephinidae Calymenidae Cyclopygidae Styginidae Homalonotidae Trinucleidae Dimeropygidae Raphiophoridae Ityophoridae Proetoidea Rorringtoniidae Acastidae Aulacopleuridae Odontopleuridae Scharyiidae Lichidae Illaenidae Brachymetop. Isocolidae Cheiruridae Pterygometop. Encrinuridae Dalmanitidae Harpetidae Phacopidae

Number of Genera 20

10

0.2

S5

20

10

20

S4

10

S3

20

10

S2

10

20

S1

10

O9

20

10

O8

20

10

O7

20

10

O6

40

30

20

O5

10

30

20

10

O4

40

30

20

O3

10

20

10

O2

20

10

O1

1.0

0.8

0.6

0.4

0

IBEX I IBEX II WHITEROCK

FIGURE 24.2. Cluster analysis of Ordovician trilobite families, with plots of their diversity through time. Clustering was based on Ordovician diversity only (intervals O1–O9). Taxa were clustered using as variables the number of genera in each of the nine Ordovician biostratigraphic intervals. The Pearson product-moment correlation coefficient was used as the index of similarity, and the clusters were formed using the average linkage method.

Trilobites radiation of megistaspidine asaphids ( Jaanusson 1953a, 1953b, 1956; Tjernvik 1956; Tjernvik and Johansson 1980; Nielsen 1995). Other elements of Baltic Arenig/Llanvirn faunas are shared with Laurentia (Celmus, Nileus, Raymondaspis, Illaenus), with which there are clearly the strongest faunal ties. Thirteen major Whiterock Fauna clades were present in the Megistaspidine Realm at the onset of radiation (table 24.1), but none were endemic to the realm. A temperate southern Dalmanitoidean Realm (Selenopeltis Province of Whittington and Hughes 1972) includes Avalonia and parts of Gondwana (e.g., Armorica, Perunica, present-day North Africa). In keeping with its high-latitude position, this area has the fewest links with the other major realms and is dominated by cyclopygids, ogygiocarinine asaphids, trinucleids, reedocalymenine calymenids, and early dalmanitids (e.g., Hammann 1974, 1983; Fortey and Owens 1978, 1987; Henry 1980; Rabano 1990). Twelve major Whiterock Fauna clades were present in the Dalmanitoidean Realm at the onset of radiation (table 24.1), of which four were endemic. A Reedocalymenine Realm also has tropical/ equatorial distribution and comprises part of Gondwana, including South China, Australia, and much of South America (Asaphopsis Province of Whittington and Hughes 1972). The Hungaiidae (= Dikelokephalinidae; see Ludvigsen and Westrop in Ludvigsen et al. 1989 for discussion) is restricted to this

2/0 13/2 13/0

12/4 Bathyurid Realm

Reedocalymenine Realm

Megistaspidine Realm

Dalmanitoidean Realm

FIGURE 24.3. Four biogeographic realms during the time of the Ordovician Radiation, as defined by Whittington and Hughes (1972). Kazakhstan and North China are left unshaded because their affinities are not definite, but they may belong to the Bathyurid Realm. Numbers indicate the number of Whiterock Fauna clades occurring in the realm at the time of the radiation and the number endemic to that realm (see table 24.1).

and cybelopsine pliomerids, though other taxa, such as dimeropygids and hystricurids, have their distributions concentrated in this realm (e.g., Ross 1951; Hintze 1953; Whittington 1963, 1965; Fortey 1979, 1980; Adrain and Fortey 1997). Thirteen major Whiterock Fauna clades were present in the Bathyurid Realm (table 24.1) at the time of onset of the Ordovician Radiation, of which two were endemic. A southern midlatitude Megistapidine Realm occupies Baltica (Asaphid Province of Whittington and Hughes 1972) and is marked by a striking endemic

TABLE 24.1. Distribution by Faunal Realm of Whiterock Fauna Families and Subfamilies During the Onset of the Ordovician Radiation (intervals O3 and O4) Taxon Calymenidae: Calymeninae Colpocorpyinae Reedocalymeninae Cheiruridae Cyclopygidae Dalmanitidae Dimeropygidae Encrinuridae Homalonotidae Illaenidae Isocolidae Lichidae Odontopleuridae: Ceratocephalinae Selenopeltinae Proetoidea Pterygometopidae Raphiophoridae Styginidae Trinucleidae

Bathyurid

Megistaspidine

Dalmanitoidean

√

√

√

√ √

√ √

√ √

√ √ √ √ √ √

√ √ √

√ √ √

√ √ √ √ √ √

√ √ √ √ √

√ √ √

Reedocalymenine

√

√ √ √

√

235

236

        .           . realm during the Arenig. Benthic platform faunas are highly endemic (e.g., Harrington and Leanza 1957; Fortey and Shergold 1984; Jell and Stait 1985b; Laurie and Shergold 1996a, 1996b). Pelagic telephinids, however (Cocks and Fortey 1990: figure 3), are shared with equatorial Laurentia and may be difficult to distinguish between continents even at the species level (Fortey 1975b; McCormick and Fortey 1999). Strikingly, only two Whiterock Fauna families, Raphiophoridae and Illaenidae, were present in the Reedocalymenine Realm during the onset of radiation. These are also the only two Whiterock Fauna groups with a global, fully cosmopolitan distribution during this time. There is therefore, for reasons thus far unknown, no evidence for a significant Ordovician radiation of trilobites in the Reedocalymenine Realm, and it was not until considerably after the radiation elsewhere (Edgecombe, Webby and Laurie, later in this chapter) that many Whiterock Fauna groups appeared in Australia. We therefore exclude this realm from consideration and concentrate on latitudinal patterns between the remaining three, which were positioned at low, intermediate, and high latitudes at the time of the radiation.

High- versus Low-Latitude Radiation and End Ordovician Extinction Trilobites are an exemplar taxon for groups hard hit by the end Ordovician mass extinction. By all estimates, including our own (figure 24.1), trilobites lost around half of their global taxic diversity during the event. The question of selectivity at a major mass extinction is always of interest. Are extinguished groups related, and different from survivors, in some particular trait or pattern? Chatterton and Speyer (1989), for example, claimed that the larvae of some trilobites were benthic, whereas others were pelagic, and that groups with the latter suffered greater extinction. Adrain et al. (1998) related extinction propensity to clade size, showing that the end Ordovician event preferentially removed clades whose latest Ordovician genus diversity was low. Is there any geographic component to end Ordovician extinction patterns? For Ordovician trilobites, geographic patterns are masked by a burst of cosmopolitanism during the

TABLE 24.2. Latitudinal Distribution of Whiterock Fauna Families and Subfamilies During the Time of the Radiation Contrasted with End Ordovician Fate Early Distribution a Taxon Calymenidae: Calymeninae Colpocoryphinae Reedocalymeninae Cyclopygidae Styginidae Homalonotidae Trinucleidae Dimeropygidae Raphiophoridae Proetoidea Odontopleuridae Lichidae Illaenidae Isocolidae Cheiruridae: Cheirurinae Survival Acanthoparyphinae Cyrtometopinae Deiphoninae Eccoptochilinae Sphaerexochinae Pterygometopidae Encrinuridae Dalmanitidae

Low

Middle

High

End Ordovician Survival EXTINCT EXTINCT EXTINCT Survival Survival EXTINCT EXTINCT Survival Survival Survival Survival Survival EXTINCT

50 0 6 4 36 0 14 81 68 60 31 20 49 45

10 11 29 13 60 15 26 19 19 0 26 55 32 45

40 89 65 83 4 85 60 0 13 40 43 25 19 10

63

26

11

0 0 75 0 94 23 62 0

100 100 25 26 6 73 19 0

0 0 0 74 0 4 19 100

Survival EXTINCT Survival EXTINCT Survival Survival Survival Survival

a This early distribution is calculated as a percentage of total species present during intervals 03, 04, and 05, at low, middle, and high latitudes.

Ashgill in response to climatic cooling. During this time, many groups that had shown strong highlatitude endemicity achieved wider, low-latitude distributions. For example, dalmanitids, chasmopine pterygometopids, reedocalymenine calymenids, homalonotids, and cyclopygids, among others, became widespread just prior to the extinction. Based on the summary presented here, Ordovician radiations of trilobites clearly occurred with a substantial latitudinal component (table 24.2). Some radiating groups were initially endemic or nearly endemic to high-latitude Gondwana and others to low-latitude Laurentia (and possibly Siberia/Kolyma, though data are very sparse). This tabulation demonstrates that a strong majority (12 of 15) of clades that had their origin and early diversification centered in the Bathyurid Realm or Megistaspidine Realm (i.e., had a majority or plurality of their species diversity occurring there) went on to survive the end Ordovician mass extinction and contribute to the Silurian Fauna. Further, of the 15 clades centered in either the Bathyurid or Megistaspidine realm, 14 are next most common in the

Trilobites other of these realms—that is, they have an extremely strong or exclusive intermediate and low-latitude distribution. The only exception is Encrinuridae, reflecting the high-latitude diversification of the dindymenines. In contrast, of eight clades that had their origin and early diversification centered in the Dalmanitoidean Realm, only three survived the end Ordovician mass extinction. The difference in survival between clades whose diversification occurred in low to intermediate latitudes and those with origins in high latitudes is statistically significant. A G-test with Williams’s correction for small sample sizes (Sokal and Rohlf 1981) rejected the null hypothesis of independence of survival from latitudinal distribution at the .05 level. Hence, for whatever reason, a significant majority of clades that would survive the extinction and form the Silurian Fauna had their first occurrence and their radiation heavily concentrated in intermediate and low latitudes, and most were centered in the Bathyurid Realm.

Development of the Whiterock Fauna in Laurentia It is important to test for any environmental signal to the emergence of the Whiterock Fauna in all three realms in which a significant radiation of trilobites occurred. However, adequate data are currently limited to the low-latitude Bathyurid Realm, which contributed more than half of the clades destined to form the Silurian Fauna. Here, the focus is on the environmental pattern of the radiation in Laurentia because this continent contributes the overwhelming amount of data for the Bathyurid Realm and also has the best available environmental information. Concern for Ordovician environmental distributions and their potential overprint on geographic patterns was pioneered by Fortey (1975a) in his qualitative analysis of Spitsbergen faunas. Ludvigsen (1978) applied quantitative techniques to biofacies analysis in the Middle Ordovician of northwestern Canada, and Q- and R-mode hierarchical cluster analysis of trilobite biofacies is now becoming routine (e.g., Ludvigsen and Westrop 1983; Westrop 1986; Ludvigsen et al. 1989; Westrop and Cuggy 1999). Fortey’s biofacies scheme is intuitive and has been widely supported, but it has not been put on a formal analytical footing, a task that we attempt here.

Ideally, biofacies analysis should be undertaken with quantitative sampling data, so that relative abundance of taxa can participate in clustering. Such data are only occasionally available (summarized on a global scale by Westrop and Adrain [1998] and Adrain et al. [2000]) and are lacking for most well-described faunas from Laurentia at the time of onset of the Ordovician Radiation. As a result, simple presence/absence data were used, with the acknowledgment that information on relative abundance may in the future greatly enhance the analysis. The results from analysis of 37 Laurentian collections from the O4 interval are shown in figure 24.4. Most taxa are genera, although occasionally subfamilies or even families are used where data were sparse. “Singleton” taxa restricted to a single collection were not scored. Analyses were performed with SPSS v.10.0 for the Macintosh computer (SPSS 2000). The analysis confirms the general applicability of Fortey’s (1975a) scheme, yielding four progressively deeperwater biofacies labeled (following Fortey) Bathyurid, Illaenid-Cheirurid, Nileid, and Olenid. The Bathyurid biofacies includes collections from shallow subtidal environments (i.e., above-average storm wavebase) common on the craton interior. The Illaenid-Cheirurid biofacies occurs in deep subtidal and buildup collections, mostly derived from the craton margin. The Nileid and Olenid biofacies occur in progressively deeper-water, mostly marginal environments. Does the Whiterock Fauna contribute more substantially to one or more of these biofacies? As demonstrated by figure 24.5, the majority of species occurring in the Illaenid-Cheirurid biofacies belong to Whiterock Fauna groups, whereas the Ibex faunas dominate in all other biofacies. The pattern is more striking still when only the Silurian Fauna (excluding the Raphiophoridae, as in Adrain et al. 1998) is plotted—groups that survived the end Ordovician were overwhelmingly concentrated in the IllaenidCheirurid biofacies. Hence, the trilobite radiation in Laurentia was initiated in marginal deep subtidal and buildup environments fringing the craton. The pattern of origin and spread of the Whiterock Fauna in Laurentia is shown in figure 24.6. Early Ibexian faunas were completely dominated by elements of the Ibex faunas across the environmental spectrum. By the late Ibexian, the Whiterock Fauna had appeared in all environments but dominated

237

0

Similarity

Bathyurid

IllaenidCheirurid

Nileid Olenid

Meik Base Juab Fm. Dounans Lst. Wahwah Fm. Eleanor River Fm. Catoche Fm. Meik Bioherm 3 Whiterock Canyon Pyramid Peak Meik Bioherm 1 Meik Bioherm 2 Tourmakeady Lst. Shallow Bay Fm. Spitsbergen V3b-3 Spitsbergen V3b-2 Spitsbergen V3b-1 Spitsbergen V4a-1 Spitsbergen V4a-5 Spitsbergen V4a-4 Spitsbergen V4a-3 Spitsbergen V4a-2 Little Rawhide Mt. Spitsbergen V3a-3 Spitsbergen V3a-2 Spitsbergen V1c-3 Spitsbergen V2a-3 Spitsbergen V2a-2 Spitsbergen V2a-1 Spitsbergen V2b-2 Spitsbergen V2b-1 Spitsbergen V3a-1 Spitsbergen V2b-3 Spitsbergen V1b-3 Spitsbergen V1c-2 Spitsbergen V1b-5 Spitsbergen V1c-1 Spitsbergen V1b-4

Jeffersonia Strigigenalis Isotelinae Cybelopsis Benthamaspis Presbynileus Punka Uromystrum Colobinion Petigurus Psephosthenaspis Pseudomera Madaraspis Heliomerinae Scotoharpes Ceraurinella Odontopleuridae Isocolus Proetoidea Celmus Glaphurus "Catillicephalidae" Calymeninae Remopleurididae Nileus Styginidae Illaenus Kawina Acidiphorus Ischyrotoma Amphilichas Ectenonotus Cybelurus Peraspis Globampyx Poronileus Evropeites Ampyxoides Shumardiidae Endymionia Xystocrania Balnibarbi Telephina Opipeuter Agnostoidea Megalaspides Mendolaspis Gog Pytine Symphysurus Niobe Ampyx Falanaspis Svalbardites Lascorsalina Carolinites Triarthrinae Hypermecaspis Cloacaspis 1.0

Similarity

1.0

0

FIGURE 24.4. Definition of Laurentian trilobite biofacies during the onset of the Ordovician Radiation. Data matrix with horizons in Q-mode clustering order and taxa in R-mode clustering order. Data are presence/absence. Jaccard’s coefficient was used as the index of similarity, and the clusters were formed using the average linkage method. Biofacies are defined using the intersection of Q- and R-mode clusters. Catoche Formation, western Newfoundland (late Ibexian); Dounans Limestone, Highland Border Complex, Scotland (late Ibexian); Eleanor River Formation, Ellesmere Island, Canadian Arctic (late Ibexian); Juab Formation, Ibex area, western Utah (early

Trilobites Percentage

80

80

A

60 40

40

20

20

0

B

I-C

N

O

B

60

0

Biofacies

B

I-C

N

O

Biofacies

FIGURE 24.5. A, percentage of total species occurrence in each early Whiterockian Laurentian biofacies that is contributed by the Whiterock Fauna (B = Bathyurid Biofacies; I-C = IllaenidCheirurid Biofacies; N = Nileid Biofacies; O = Olenid Biofacies). B, percentage of total species occurrence contributed by the Silurian Fauna, excluding Raphiophoridae (see Adrain et al. 1998).

none. Transition to dominance by the Whiterock Fauna first occurred in deep subtidal and buildup environments during the early Whiterockian—at a time when the Ordovician Radiation of the Paleozoic Evolutionary Fauna was taking place. The Whiterock Fauna subsequently spread both on- and offshore, dominating all environments by the Late Ordovician and of course completely dominating Silurian environments with the extinction of all Ibex faunal elements by the end Ordovician.

Conclusions and Prospects from the Global Analysis 1. Ordovician trilobite families display one of three major diversification trajectories: earliest Ordovician success followed by steady and rapid decline and extinction prior to or at the end Ordovician (Ibex Fauna I); Early Ordovician radiation followed by Mid Ordovician decline and extinction prior to or at the end Ordovician (Ibex Fauna II); or Mid Ordovi-

FIGURE 24.4. (Continued ) Whiterockian); Little Rawhide Mountain, Antelope Valley Limestone, central Nevada (late Ibexian); Meik Base = Meiklejohn Bioherm basal beds (early Whiterockian), southwestern Nevada; Meik Bioherm 1–3 = Meiklejohn Bioherm, separate collections from biohermal beds (early Whiterockian); Pyramid Peak, Death Valley, California (early Whiterockian); Shallow Bay Formation, western Newfoundland (early Whiterockian); Spitsbergen V1–V4, various stratigraphic levels in the Valhallfonna Formation, drawn from Fortey (1980: figure 1), ranging from late Ibexian to early Whiterockian; Tourmakeady Limestone, western Ireland (early Whiterockian); Wahwah Formation, Ibex area, western Utah (late Ibexian); Whiterock Canyon, Antelope Valley Limestone, eastcentral Nevada (early Whiterockian).

cian radiation and Late Ordovician success, with substantial survivorship at the end Ordovician (Whiterock Fauna). 2. There was an Ordovician radiation of trilobites in all parts of the world except the tropical Gondwanan Reedocalymenine Realm. The radiation occurred at the same time as, and with dynamics similar to, that of the Paleozoic Evolutionary Fauna. 3. Whiterock Fauna clades that diversified at high latitudes were less likely to survive the end Ordovician mass extinction, whereas almost all of those that diversified at low latitudes survived. 4. In Laurentia, the Whiterock Fauna rose to dominance first in marginal deep subtidal and buildup environments and spread later to progressively dominate more on- and offshore environments. Future research could be focused on improving all the analyses with more and better data and completely stepping the level of global resolution down to species level. Greater global stratigraphic resolution using trilobites seems unlikely in the foreseeable future owing to the limited study currently being undertaken in major areas. However, further progress could certainly be made in continents with good records. Particular questions requiring further investigation include the following: 1. What is the biofacies pattern of the transition in the Megistaspidine and Dalmanitoidean realms? Relative abundance data are even more sparse than for Laurentia (see Adrain et al. 2000), but it should be possible to collect presence/absence data. 2. What does the marginal Laurentian emergence of the Whiterock Fauna signify? Is there a major extrinsic cause for the trilobite radiation? 3. Are the radiations in the three realms one, two, or three events? The high-latitude radiation in the Dalmanitoidean Realm seems to precede that in the Bathyurid Realm (Owen and McCormick 2003). Did it have a different cause? ■

Regional Patterns in Australia and New Zealand (GDE, BDW, JRL)

Ordovician trilobite faunas of Australasia were reviewed by Webby and Edgecombe in Webby et al. (2000), with particular reference to their biogeographic affinities. Most of the data in that synthesis,

239

240

        .           . FIGURE 24.6. Representative Laurentian faunas through time and along an environmental gradient, showing the pattern of changeover from the Ibex faunas (black) to the Whiterock Fauna (white), with number of species recorded from each fauna.

Nearshore

Faunas unknown

Faunas unknown

Shallow Subtidal

Buildups

Deep Subtidal

Slope

S=6

S = 15

S = 43

S=2

Rochester Fm., New York State

Drommebjerg Lst., northeastern Greenland

Cape Phillips Fm., Arctic Canada

Cape Phillips Fm., Arctic Canada

S = 19

S = 50

Irene Bay Fm., Arctic Canada

Faunas unknown

S = 17

Chair of Kildare Lst., Ireland

S=9

S = 15

S=4

Cincinnatian

Whittaker Fm., Northwest Territories

Frankfort Shale, New York State

S = 16

S=4

Mohawkian Verulam Fm., southern Ontario

Effna Fm., Virginia

Whittaker Fm., Northwest Territories

S = 13

S = 18

S = 23

Sunblood Fm., Northwest Territories

Bromide Fm., Oklahoma

Chazy Grp. reef facies, New York State

Esbataottine Fm., Northwest Territories

S = 43

S = 34

S=6

S = 11

Edinburgh Fm., Virginia

S=5

S=4

Sunblood Fm., Northwest Territories

Silurian

unnamed carbonate unit, Yukon Territory

S=8

Juab Fm., Utah

Shallow Bay Fm., Newfoundland

S = 24

S = 16

S = 18

Catoche Fm., Newfoundland

Wahwah Fm., Utah

Shallow Bay Fm., Newfoundland

Al Rose Fm., California

S=9

S=7

S=4

S=4

S=3

Boat Harbour Fm., Newfoundland

House Limestone, Utah

Shallow Bay Fm., Newfoundland

Goodwin Fm., Nevada

Cow Head Group, Newfoundland

as well as the present analysis, are from Australia (see Webby and Nicoll 1989: appendix 2 for a species list). New Zealand is represented only by Lower Ordovician faunas of the Patriarch Formation and Summit Limestone in the Takaka Terrane, South Island (Wright et al. 1994) and a few species from the Upper Ordovician Golden Bay Group of the adjacent Buller Terrane. The trilobite record in Australia samples a broad range of biofacies. For example, Lancefieldian/ Tremadocian assemblages include platform sediments representing the inner detrital belt (Shergold 1991) and peritidal carbonates (Shergold 1975), as well as an outer detrital biofacies sampled in western New

Table Cove Fm., Newfoundland

S=5

upper Whiterockian

lower Whiterockian

Valhallfona Fm., Spitsbergen

S = 16

upper Ibexian

Valhallfonna Fm., Spitsbergen

lower Ibexian

South Wales (Webby et al. 1988) and New Zealand (Wright et al. 1994). A succession of Ordovician faunas outlined by Webby et al. (2000) drew heavily on the biogeographic sensitivity of trilobites. The Lower Ordovician in Australasia was divided (in ascending order) into Hysterolenus, Australoharpes, Koraipsis, Chosenia, and Encrinurella faunas, all having affinities to North China and most having affinities to the Sibumasu Terrane. Relationships to China and Sibumasu are maintained through the Middle Ordovician Railtonella and Prosopiscus faunas, the latter also having links to Laurentia and the Precordillera of Argentina. A Gisbornian Incaia fauna in New Zealand

Generic Turnover (Rates per million years)

Trilobites FIGURE 24.7. Generic turnover (appearance and disappearance rate per million years) and normalized generic diversity curves for Ordovician trilobites of Australia and New Zealand. Normalized diversity is plotted for sampled and range-through data. Be = Bendigonian; Ch = Chewtonian; Cast = Castlemanian; Yap = Yapeenian.

12

GENERIC TURNOVER: (a) Appearances

10

x


(b) Disappearances 8

x


6

x


x
4


x


x


x


x

x
x


x



x


x

2


x x x


x


0

x


x



x

32

Normalized Diversity (Generic)

GENERIC DIVERSITY TOTALS: (a) Normalized diversity (sampled data) (b) Normalized diversity (range-through data)

24

16




x
8

0

1a

1d 1b 1c Lancefieldian

2a

2b Be

TREMADOCIAN

LOWER ORDOVICIAN 489 Ma

4c 2c 3a 3b 4a 4b Ch Cast/ Yap Darriwilian

5a 5b Gisbornian

5c 5d Eastonian

6a 6b 6c Bolindian

DARRIWILIAN

MIDDLE ORDOVICIAN

UPPER ORDOVICIAN

460.5

has South American affinities, whereas EastonianBolindian trilobites of Australia (EokosovopeltisPliomerina fauna) display a clear association with Kazakhstan, North China, and terranes near the margin of Gondwana. Elements of the cosmopolitan Hirnantia Fauna (brachiopods and trilobites) have been recorded in Victoria and Tasmania. Diversity data for Australia are affected by biases in geographic representation through the Ordovician. The Canning Basin in Western Australia (Legg 1976; Laurie and Shergold 1996a, 1996b) has a full record for the Early and Mid Ordovician but was a site of nonmarine and evaporite sedimentation in the Late Ordovician. Late Ordovician data are derived largely from the island-arc complexes of central-west New South Wales and platform carbonate sequence of Tasmania. Most known Ordovician faunas in Australasia have been described, and we have included unpublished data for the Amadeus Basin, Tasmania, and central-west New South Wales to eliminate bias. The diversity curves in figure 24.7 have been compiled at the generic level, but because a majority of genera are represented by single species, discrepancies between generic and specific diversity are not a major issue. The relatively short temporal duration of many genera is reflected in a general correspondence be-

443

tween appearances and disappearances through most of the Ordovician, and longer-ranging genera appear to be randomly distributed through time. Intervals in which disappearances substantially exceed appearances are TS.4c (late Darriwilian) and TS.6a (early Bolindian). Qing et al. (1998), Shields and Veizer (chapter 6), and Barnes (chapter 8) have demonstrated a dramatic decline in strontium isotope ratios across the late Darriwilian to early Caradoc (TS.4c–5a) interval. Qing et al. (1998) and Barnes link this change to a possible global mantle plume event with associated orogenic quiescence and/or major flooding (the early Caradoc transgression of Cocks and Fortey 1988). Shields and Veizer attribute the cause to increased ocean-ridge spreading rates. In contrast, the TS.6a decline seems to have been related more to regional than to global factors, with the Benambran Orogeny causing major tectonic changes and volcanic activity, especially to eastern Australia (Webby and Percival in Webby et al. 2000:109). Range-through and sampled diversity both show four maxima (listed in descending order using the range-through data) in the Bendigonian (TS.2b), Eastonian (TS.5d), Darriwilian (TS.4b-4c), and Lancefieldian (TS.1b). Range-through diversity particularly exceeds sampled diversity through two intervals of

241

242

        .           . the Ordovician, TS.1c–2a and TS.3a–4a. The latter (Castlemainian–early Darriwilian) is one of three marked minima in diversity by either measure, the others being in the Gisbornian (TS.5a–b) and in the globally depauperate late Bolindian (TS.6b–c). The diversity low through the TS.3a–4a interval may be at least in part an artifact of incomplete sampling, limited exposure, and/or restriction of favorable trilobitebearing facies (Georgina Basin and Tasmania), and/or breaks (disconformities) that represent localized emergence through a part of this TS.3a–4a interval (Canning Basin). Adrain et al. (earlier in this chapter) observed geographic variation in the expansion of their Whiterock Fauna (Adrain et al. 1998). Owen and McCormick (later in this chapter) noted an early (early Arenig) diversification of elements of this fauna in Avalonia, whereas it diversified relatively late in South America (Waisfeld, later in this chapter). Lower Ordovician faunas representing the inner detrital biofacies throughout Australia (e.g., Jell 1985; Jell and Stait 1985a, 1985b; Shergold 1991) are composed almost exclusively of members of the Ibex Fauna (e.g., asaphids, pilekiids, pliomerids, and hystricurids), like that of the diverse, highly endemic Bendigonian succession of the Canning Basin, where asaphids, pliomerids, and telephinids predominate (Laurie and Shergold 1996a, 1996b). In the Canning Basin, elements of the Whiterock Fauna such as illaenids and raphiophorids first rival the diversity of the Ibex Fauna in the Darriwilian (TS.4b–c; Legg 1976). Whiterock taxa, notably encrinurids, cheirurids, lichids, styginids, illaenids, and trinucleids, become dominant for the first time in the Eastonian faunas of New South Wales and Tasmania. Interval TS.5c (early Eastonian) is marked by generic appearances greatly exceeding disappearances, which appears to correspond to the expansion of the Whiterock Fauna. Ibex elements (shumardiids, agnostids, telephinids) remain the sole or dominant component only in some Late Ordovician outer detrital biofacies (e.g., Shoemaker Beds in Tasmania: Burrett et al. 1983; Oakdale Formation in New South Wales: Webby 1974). This is, interestingly, opposite to the pattern for South America (Waisfeld, later in this chapter), where Ibex taxa preferentially survive in shallowwater settings. The near absence in the Australian record of particular shallow-water Ibex groups such as

bathyurids may figure in the difference between the South American and Australian patterns. The domination of the Whiterock Fauna in Australia is even later than in South America. Although the paucity of early Mid Ordovician (Castlemainian–early Darriwilian) trilobite faunas is a problem for understanding this transition, even later Darriwilian asemblages (e.g., Nora Formation in the Georgina Basin: Fortey and Shergold 1984) are composed almost exclusively of Ibex taxa such as asaphids, leiostegiids, dikelokephalinids, telephinids, and prosopiscids. ■

Regional Patterns in South America (BGW)

Trilobite records in the Ordovician of South America are derived from three different geodynamic and environmental settings: pericratonic platforms developed along the southwestern edge of Gondwana (Andean belt), volcanic island-arc settings parautochthonous to the Gondwanan margin (Famatina Range and western Puna), and an allochthonous Laurentianderived terrane (Argentine Precordillera). The Andean belt records deposition in shallow marine siliciclastic shelves. In northern South America, Ordovician deposits are poorly known, and trilobite records are sparse. In contrast, better-known and trilobite-rich successions, mostly early Tremadocian to mid Arenig in age, crop out in the Central Andean basin (Cordillera Oriental of Peru, Bolivia, and northwestern Argentina). Trilobite species diversity is high in the early Tremadocian to mid Arenig (figure 24.8). Tremadocian successions are composed of transgressive deposits that frequently represent dysaerobic environments, punctuated by regressive episodes (cf. Moya 1988) represented by intertidal and shallow subtidal settings. The greatest number of species occurs in the early Tremadocian, associated with the global rise in sea level and the development of muddominated platforms, dominated by relatively widespread olenids and several families of agnostoids. High diversity was maintained in the late Tremadocian to the early and mid Arenig, with a remarkable radiation of endemic asaphids in the early Arenig (T. approximatus to B. deflexus graptolite zones) associated with dysaerobic facies in distal parts of the inner shelf (Waisfeld et al. 1999). In the mid Arenig (uppermost B. deflexus and D. bifidus graptolite zones),

Trilobites FIGURE 24.8. Normalized diversity of trilobite species in South America. A, aggregate diversity curve for all the basins. B, diversity curve for the Andean belt. C, diversity curve for Famatina Ranges and Western Puna. D, diversity curve for Argentine Precordillera.

in shallower and more aerobic waters of the inner shelf, there was a radiation of endemic trinucleids, raphiophorids, pliomerids, nileids, and asaphids. Diversity increase is also related to the occurrence of a few immigrants from northwestern Gondwana (Waisfeld 1995). Post-Arenig trilobite records in the Andean belt are sparse and restricted to isolated outcrops and, except in Argentina, mostly lack accurate chronological constraints. A few asaphids, along with calymenids, homalonotids, and trinucleids, are recorded in Llanvirn to Caradoc strata. In Argentina there is evidence for active vulcanism during the mid Arenig in the Famatina Range and during the early Tremadocian and the mid Arenig in the western Puna region (Astini and Benedetto 1996; Kouhkarsky et al. 1996; Mángano and Buatois 1996). Trilobite information is still preliminary, particularly in the latter region. In the Famatina Range, trilobites exhibit the highest diversity in siliciclastic and volcaniclastic deposits of the proximal- to middle-shelf environment, associated with an increase in volcanic activity (conodont-based O. evae Zone). No particular radiation of endemic forms is recognized in this basin, but immigrants from West Gondwana coexist with East Gondwanan and Baltic forms (Vaccari 1995; Waisfeld and Vaccari 1996). The Argentine Precordillera terrane rifted from Laurentia in the Early Cambrian and drifted from low to high latitudes during the Ordovician (Astini et al. 1995). From the latest Tremadocian, open marine carbonates prevailed, yielding a relatively complete record of the benthic fauna. Late Tremadocian to early Llanvirn carbonate sedimentation is represented by inner- and middle-ramp settings (Cañas 1999). Trilobite diversity is relatively low and composed largely of representatives of the Ibex Fauna.

The base of the Whiterockian is within this carbonate succession, but it is not marked by particular expansion of families belonging to the Whiterock Fauna. Endemic genera are restricted to a single pliomerid in the mid Arenig and a single scutelluinid in the early Llanvirn (Vaccari 2003) both from the middle ramp. Both are associated with immigrants from the lowand intermediate-latitude Bathyurid Province of East Gondwana and Baltica (cf. Vaccari 1995). A change from the middle to distal ramp and a shift in facies from exclusively carbonate to mixed carbonate-siliciclastic sedimentation took place diachronously from late in the mid Arenig to the early Llanvirn. This facies change is linked to a tectonic shift associated with the generation of subsiding depocenters and with global sea level rise in the early Llanvirn (mid Darriwilian) (Astini 1999a; Cañas 1999). This shift was critical in the expansion of trilobites. The earliest records (conodont-based B. navis to M. parva zones) of trilobites in the distal ramp environment comprise relatively widespread elements (nileids, olenids, raphiophorids, etc.), but in the early Llanvirn (conodont-based E. suecicus Zone) these are associated with a radiation of new genera, particularly Raphiophoridae, and also Trinucleidae and Toernquistiidae. Families belonging to the Whiterock Fauna account for 40 percent of the fauna. Caradoc and younger deposits show a strong facies differentiation in the Precordillera. Trilobites are particularly restricted to carbonate remnants that persisted locally after carbonate sedimentation was drowned in most parts of the basin. Their occurrence and diversification appear to be strongly controlled by environmental constraints imposed by the regional topography and local tectonics. An early Caradoc (graptolitebased N. gracilis and C. bicornis zones) peak in trilobite

243

244

        .           . diversity is coincident with one of these carbonate remnants. Trilobites occur in slope-apron deposits in either autochthonous deep-water limestones (hemipelagites) or slightly shallower water resedimented carbonates (cf. Astini 1995). New species of earlier endemic raphiophorids and toernquistiids occur, together with endemic encrinurids and a remarkable radiation of trinucleids. The latitudinal position of the Precordillera during the Caradoc is still debated, with mixed biogeographic affinities of the fauna indicating proximity to the Gondwanan margin and a possible location at intermediate latitude (Benedetto 1998; Benedetto et al. 1999). Trilobite records in the late Caradoc to Ashgill of the Precordillera are scarce, with only a few homalonotids and dalmanitids associated with the Hirnantia Fauna. Biodiversification of trilobites in South America shows a strong biogeographic and environmental overprint. In the Andean belt, local radiation of forms is remarkable in the early and mid Arenig among the representatives of the Ibex Fauna. The Whiterock Fauna is well developed in the Precordillera, in contrast to the Andean belt, where Middle and Upper Ordovician records are limited. The Whiterock Fauna diversified diachronously from late in the mid Arenig to the early Llanvirn, tracking the progressive development of distal ramp settings in the basin, while trilobite families of coeval shallower settings are still dominated by representatives of the Ibex Fauna. This pattern is similar to that reported by Adrain et al. (1998) in Laurentia. However, the initial development of the Whiterock Fauna appeared to take place slightly later. This could be a result of the appropriate distal facies being developed slightly later in the basin. ■

Regional Patterns in the Anglo-Welsh Sector of Avalonia (AWO, TMcC)

The Anglo-Welsh area contains the most complete Ordovician successions from Avalonia. They record the history of that microcontinental terrane from its Early Ordovician location on the intermediate to high-latitude Gondwanan margin, probably close to West Africa (McNamara et al. 2001), through its early Mid Ordovician rifting and northward drift leading to its collision with Baltica in the Late Ordovician and the Laurentian margin in the Early Silurian (Cocks et

al. 1997; van Staal et al. 1998; Cocks 2000). The Anglo-Welsh Ordovician trilobite faunas are well documented, and their temporal and spatial diversity patterns have been investigated using a literaturebased relational database in which the occurrences of species at localities are linked to an array of taxonomic, geographic, and stratigraphic data (Owen and McCormick 1999; McCormick and Owen 2001). We have described elsewhere the patterns of trilobite diversity change in the Ordovician of the Welsh basin at genus and species level (McCormick and Owen 2001) and the whole Anglo-Welsh area at genus level (Owen and McCormick 2003). In doing so, we have demonstrated that members of the Whiterock Fauna became the dominant component of the Avalonian trilobite fauna earlier (early Arenig) than its rise to dominance in global terms and that they were fairly evenly distributed through the whole spectrum of shelf to upper-slope environments. These analyses were undertaken using the stages recognized in what is the historical type area for the Ordovician (Fortey et al. 1995, 2000) as the “time slices” and utilizing a simple measure of diversity that counted as unity the occurrence of a taxon within a given stage or inferred (in the case of “range-through” analyses) to have been present in that stage because of its presence in the preceding and succeeding intervals. We present here species- and genus-level curves (figure 24.9) using the normalized diversity measure recommended for the IGCP 410 clade analyses and equating the Anglo-Welsh stratigraphy as closely as possible to the time slices recommended for the international project (chapter 2). Few of the boundaries between the latter divisions match exactly the chronostratigraphic or biostratigraphic boundaries defining the units within which the data were compiled, and the equivalencies used are shown in the caption to figure 24.9. The normalized diversity measure per time slice counts the number of taxa occurring within the slice and also found above and below it, plus half the number of those taxa that either originated or became extinct during that interval plus half the number of taxa restricted to the slice. Two sets of curves have been computed, one based on recorded occurrences within each slice and one that infers the existence of taxa within an interval on the basis of the “range-through” principle (see earlier in this chapter). The stratigraphic age of some of

Trilobites

FIGURE 24.9. Sampled and range-through trilobite normalized biodiversity curves for the Anglo-Welsh sector of Avalonia. Note that the boundaries between many of the time slices recommended by Webby et al. (chapter 2) cannot be recognized in the Anglo-Welsh area or that the shelly faunal data could not be compiled relative to them. The time slices as used here have the following equivalencies in the Anglo-Welsh chronostratigraphy: TS.1a = Cressagian; TS.1b = early Migneintian; TS.1c = mid Migneintian; TS.1d = late Migneintian; TS.2a = early Moridunian; TS.2b = late Moridunian; TS.2c = Whitlandian; TS.3a = early Fennian; TS.3b = mid Fennian; TS.4a = late Fennian; TS.4b = early Abereiddian (= graptolite-based D. artus Zone); TS.4c = late Abereiddian + Llandeilian; TS.5a = Aurelucian; TS.5b = Burrellian; TS.5c = Cheneyan + Streffordian; TS.5d = Pusgillian; TS.6a = Cautleyan; TS.6b = Rawtheyan; TS.6c = Hirnantian.

the records in the database is known only to two or (rarely) more stages, in which case the taxa are counted as 0.25 in any resultant time slice above or below the unequivocal range of the taxa concerned and, in the case of the “sampled” data, 0.5 where the uncertain record fills in part of the known range of the taxon. The difference between the sampled and range-through curves is particularly marked in intervals where the rock succession, and therefore the sample coverage, includes only a limited set of biofacies. Moreover, because many species are confined to a single stage (e.g., Thomas et al. 1984) whereas most genera have a much longer range, the lower weighting placed on taxa restricted to a time slice in the diversity index produces the apparently anomalous situation of there seeming to be more genera than species in many time slices. Compared with our earlier, genus-level analysis (Owen and McCormick 2003), the twofold division of the Moridunian stage produces a curve showing a more even rise in diversity through the Arenig, and the combining of the graptolite-based murchisoni Zone and Llandeilian stage and the Cheneyan and Streffordian stages to comprise TS.4c and TS.5c, respectively, produces smoother “sampled” diversity curves. Otherwise, patterns emerge from the analysis

that are similar to those obtained earlier, testifying to the robustness of the signals. Bootstrap tests (multiple random resampling of the data in each time interval—see Gilinsky and Bambach 1986 for other examples) of the sample data in our earlier analyses show that (1) the number of samples or the range of biofacies preserved within a time unit can have a strong influence on the sampled diversity and (2) the range-through curves are probably a closer reflection of the true picture (McCormick and Owen 2001; Owen and McCormick 2003). The range-through diversity curves presented herein (figure 24.9) show a rise through the Arenig to a late Arenig–early Caradoc (TS.4a–5a) plateau at genus level but a slight early Llanvirn (TS.4b) peak at species level. McCormick and Owen (2001) suggested that elevated levels of species richness of genera in the late Arenig–early Llanvirn in the Welsh basin may be linked to the rifting of Avalonia from Gondwana at that time (Cocks et al. 1997). Specieslevel diversity apparently fell during the Caradoc to earliest Ashgill (TS.5a–d), but this is to some extent an artifact of a combination of the restriction in range of preserved shelf biofacies as the basin deepened and the absence of the very deepest water trilobite biofacies from the Anglo-Welsh area. The latter, the

245

246

        .           . cyclopygid-atheoptic association, was composed largely of long-ranging genera that reappeared in the Ashgill, and hence the range-through genus-level data show a trend contrary to the species data. The mid Ashgill (Cautleyan-Rawtheyan) peak in diversity in the Anglo-Welsh area contrasts markedly with the global curves (Adrain et al., earlier in this chapter), which show a maximum in the late Arenig and Llanvirn and a considerable decline thereafter. The CautleyanRawtheyan peak in both genus and species diversity and the high species-to-genus ratio reflect the extreme heterogeneity of the environment and hence the wide spectrum of trilobite biofacies preserved in the Anglo-Welsh area prior to the extinctions that led to the Hirnantian diversity crash (Owen and McCormick 2003). ■

Regional Patterns in Baltica (ØH)

Baltoscandia provides the most complete record for the biodiversity history of Baltica during the Ordovician. It is believed that the region as a whole preserves a fairly complete stratigraphic sequence, although formations can be condensed, in particular in the Baltic countries. The genus- and species-level diversity curves presented here (figure 24.10) have been derived from a larger database (openly available on the Internet at http://asaphus.uio.no), covering all major fossil groups from the Ordovician of Baltoscandia (Hammer in press). Owing to the nature of the literature, especially the older publications, it has not been possible to collect sufficient data about individual

140

Genera

samples. The basic unit of the database is therefore first and last occurrences of a species at one locality, according to one publication. As far as possible, the “first-appearance datums” (FADs) and “last-appearance datums” (LADs) are then converted to apparent, calibrated ages according to the timescale used in this volume and correlations with local zones according to recent literature. In cases in which only the formation is known, the approximate ages of the lower and upper formation boundaries are used as the FAD and LAD, potentially overestimating real ranges and also by necessity disregarding diachronous lithological boundaries. The diversity estimates within each time slice (chapter 2) are then made using the range-through assumption and the normalized diversity count, whereby taxa having their FAD and/or LAD within a time slice count as only one-half of a unit within that slice. An alternative estimate involves counting taxa restricted to a time slice as one-third of a unit. There is no correction for the different durations of the time slices. Even though sample data have not been collected, it is to some degree possible to estimate sampling coverage by constructing “artificial samples” each consisting of taxa registered within a fixed time duration (e.g., one million years) at one locality. These quasi samples can then be subjected to bootstrap tests and other randomization methods (Hammer in press). At the time of writing, the database consists of 10,340 stratigraphic ranges at localities, taken from 141 publications. Some 962 species and 259 genera of trilobites are included, with a total of 2,691 range entries. The trilobite taxonomy has been revised ac-

Species

120 100 80 60 40 20 0

1b 1c 1d 2a 2b 2c 3a 3b 4a 4b 4c Arenig Tremadoc Llanvirn Volkhov Kunda Latorp DARRIWILIAN TREMADOCIAN LOWER ORDOVICIAN M. ORDOVICIAN

1a

489 Ma

472

5a

5b 5c Caradoc Viru

5d

6a 6b 6c Ashgill Harju

UPPER ORDOVICIAN

460.5

FIGURE 24.10. Range-through, normalized trilobite diversity curves for Baltoscandia, representing Baltica.

443

Trilobites cording to Bruton et al. (1997), but the taxonomic uncertainties in the material are still rather extensive. It can only be hoped that the taxonomic problems are relatively unsystematic and will therefore deteriorate the signals, rather than producing false ones (Benton 1999). The main feature of the curves is a more or less even, substantial increase in trilobite biodiversity throughout the period. In addition to global evolutionary trends, such a pattern may also have been influenced by local factors. Baltica drifted from a mid- to highlatitude position in the Southern Hemisphere in the Early Ordovician to low latitudes by the end of the Ordovician (Torsvik et al. 1992). In the context of the present-day latitudinal diversity gradient (Rosenzweig 1995), this movement may have contributed to the increase in diversity that is observed in all Baltoscandian fossil groups through the Ordovician. However, for trilobites, the diversity increase is less clear when the data are subjected to randomization tests (Hammer in press). At the specific level, a steep increase in trilobite diversity is observed in TS.3a (mid Arenig) and is strongly supported by the randomization tests. This diversification event close to the Ibexian-Whiterockian boundary is contemporaneous with the beginning of a probably quite protracted regression in the area (Nielsen 1995:61). A diversity peak in the species curve in the upper Llanvirn (TS.4c) is mainly due to Estonian data (in particular, those of Rõõmusoks 1970) and is not observed in the curves for Norway or Sweden. Trilobite diversity reached its all-time high in the early Ashgill (TS.6a) before dropping significantly during the late Ashgill. Even taking into account the artifacts produced by not counting Lazarus taxa, this end Ordovician decrease in diversity seems as dramatic as elsewhere in the world. ■

Regional Patterns in South China (ZYZ)

The South China Block exhibits extensive exposures of Ordovician deposits, with the most complete sequences and occurrences of fossil groups in China. The strata are well documented and the stratigraphic units highly resolved, with many selected as stratotypes for classifying and correlating the Ordovician of China. As reviewed by Cocks (2001), the South

China Block was situated in low-latitude zones along the western margin of Gondwana during the Ordovician. Cocks and Torsvik (chapter 5) further considered the South China Block as a peri-Gondwanan terrane that, like the Sibumasu Terrane, drifted from intermediate to low latitudes outboard of various Himalayan fragments close to the Indian part of Gondwana. Ordovician trilobites are well recorded (e.g., Lu 1975), and faunas display a progressive on-shelf to off-shelf transition in composition and diversity from present-day west either to the southeast or to the north and northeast of the block, with benthic forms most diversified in the shallower outer shelf and mesopelagic cyclopygids mainly distributed in the deeper outer-shelf and off-shelf slope (see Zhou et al. 1999, 2001, 2003; Yuan et al. 2000). Trilobite faunas exhibit close relationships to those of Australasia, Kazakhstan, the North China and Tarim blocks, and the Sibumasu Terrane on the one hand and to the Middle East, southern, central and western Europe, and the Indochina Terrane on the other (Zhou and Dean 1989; Zhou et al. 1998a, 1998b), providing evidence of links between the different Ordovician Gondwanan and peri-Gondwanan faunas of higher to lower latitudinal zones. In the absence of an existing database of the geologic and geographic distributions of the Ordovician trilobites in the South China Block, the genus diversity curves (figure 24.11) presented here are derived from a genus-range chart that will be described more fully in a future paper. The chart was compiled on the basis of data from the literature and collections made recently by Zhiyi Zhou, Zhiqiang Zhou, and Wenwei Yuan from 36 measured Ordovician sections along a bathymetric gradient in South China. The basic data comprise 220 taxonomically valid trilobite genera. Of these the Asaphida are the predominate group, represented by up to 41.6 percent of the entire South Chinese Ordovician trilobite generic component. The diversity data are presented using the unified Ordovician timescale with subdivisions into 19 time slices and correlations to equivalent Chinese chronostratigraphic intervals as outlined by Webby et al. (chapter 2). The data were calculated using normalized diversity measures, as recommended by Cooper (chapter 4). Range-through diversity maintains a relationship roughly similar to the sampled diversity through the

247

248

        .           . FIGURE 24.11. Sampled and range-through normalized generic biodiversity curves for Ordovician trilobites of the South China Block.

Sampled Range-through

Early to Mid Ordovician (figure 24.11), but then the differences between the two sets of values become less marked during the Caradoc and eventually more or less coincide through the Ashgill (TS.6a–c interval). The overall diversity trend shows an initial sharp rise in TS.1a and then a decline to a low in TS.1d. This is followed by a rise to a broadly flattened to slightly elevated peak during the Mid Ordovician centered on TS.4a. Then a more intense radiation occurred during the Caradoc–early Ashgill (TS.5a–6a interval) with peak diversification in TS.5b, followed by a dramatic decline through the middle Ashgill to TS.6c. Although the limited TS.1a diversity increase in the early Tremadocian seems comparable to that of the South American, particularly the Andean Belt, plot (figure 26.8), the sharp diversity decrease in the Hirnantian is synchronous worldwide. The more or less steady rise in diversity shown here from the Arenig to early Ashgill is also revealed in Avalonia (figure 24.9) and Baltica (figure 24.10), especially the latter, where a similar Caradoc–early Ashgill diversity plateau also occurs. However, as a whole, the Ordovician trilobite biodiversity curves displayed in the different regions do not match one another very closely. In summary, the diversity plot of South China shows peaks in the earliest Tremadocian, early Darriwilian, and Caradoc and two minima at the end of the Tremadocian and in the Hirnantian. The mid Ashgill diversity maximum illustrated by Owen and McCormick using Avalonian data (figure 24.9) is not depicted in South China, probably because there is a paucity of trilobite-rich rocks through this interval. Mechanisms that triggered the Ordovician bio-

diversity alternation may have involved a range of geologic, geographic, climatic, and oceanographic factors (Webby 2000). With the exception of the end Ordovician glaciation that caused the extinction of Ibex Fauna I and II and the decrease in diversity of the Whiterock Fauna (Adrain, Westrop, and Fortey, earlier in this chapter), the trilobite diversity changes exhibited in South China seem related generally to the sea level fluctuations delineated by Fortey (1984) and Ross and Ross (1992). High diversities seem to be associated with transgressive phases, and low diversities coincide with regressive intervals. The Caradoc diversity maximum may, for example, be connected with the Ordovician climax of transgression that took place in China (Zhou et al. 1989, 1992), as it did elsewhere in the world (Fortey 1984). It remains to be determined whether the correspondence between sea level and diversity change represents cause and effect or whether the diversity curve includes sample biases linked to sea level change (e.g., see Smith 2001). In the South China Block, taxa of Ibex Fauna I had high diversity in TS.1a, and most of them belong to the Hysterolenus fauna of latest Cambrian–earliest Tremadocian, including mainly agnostids, dikelokephalinids, ceratopygids, kainellids, and olenids. The diversity of Ibex Fauna I declined steeply to TS.1d, and from that point on the related forms never became diversified again (figure 24.12B). The Ibex Fauna II (chiefly Asaphidae and Nileidae) peaked in TS.2b, but otherwise its diversity was uniformly low through the Ordovician (figure 24.12B); members of it were only proportionally higher in TS.1d–3a

Trilobites

Number of Genera

Percentage

FIGURE 24.12. Proportion (A) and diversity (B) of Ordovician trilobite genera belonging respectively to the Ibex I, Ibex II, and Whiterock faunas (Adrain, Fortey, and Westrop, this chapter) in each of 19 sampled intervals of the South China Block.

(figure 24.12A), suggesting that this fauna had once radiated here during the latest Tremadocian to early late Arenig. Both Ibex I and II faunas became extinct just prior to the end Ordovician mass extinction. The main radiation of the Whiterock Fauna is recorded from the Mid Ordovician onward, with the peak of diversification attained during TS.5b (figure 24.12B). A slight decline in diversity followed through to TS.6a, and then there was a rapid decrease to the Hirnantian (figure 24.12B). The fauna was proportionally dominant over that of either the Ibex Fauna I or II from TS.3b (49 percent of total fauna) to TS.6c (100 percent) (figure 24.12A), consisting of mainly cyclopygids, cheirurids, raphiophorids, trinucleids, illaenids, calymenids, and isocolids. Most of the members were outer-shelf dwellers. The Whiterock Fauna survived the two-phase end Ordovician mass extinction and, after a short period of recovery,

reappeared in South China as the Silurian Fauna with representatives of up to 11 of its families during the mid Llandovery (mid to late Aeronian). In South China the main diversification of the Whiterock Fauna seems to have commenced during TS.3b, that is, early in the Mid Ordovician (just prior to the Darriwilian), when elements of the fauna occupied most environmental niches for the first time. This included the diversification of the mesopelagic cyclopygids as the Cyclopygid biofacies (see Fortey, later in this chapter) became established in South China during TS.3b. ■

Adaptive Deployment (RAF)

Biofacies profiles with their accompanying suites of trilobite faunas have now been recognized in the Ordovician for all the major paleocontinents. Although these follow a broadly shallow- to deep-water

249

250

        .           . trajectory, the factors controlling their distribution may be only secondarily related to depth per se—for example, the level of oxygenation present at the sediment surface may be the prime influence on which organisms are present in any given location, and characteristic trilobites adapted to deoxygenated habitats may have a depth spread. The range of niches occupied by the trilobites is an important part of the Ordovician Radiation. The taxonomic constitution and guild distribution of trilobite “communities” during the Ordovician are considered here along with the ways in which they differed from those in the Cambrian and Silurian. In so doing, assumptions must be made about life habits, which are by no means universally agreed. Fortey and Owens (1990a) identified a series of typical morphologies (“morphotypes”) repeatedly adopted by trilobites that subsequently (Fortey and Owens 1999) was extended to recognize feeding habits and habitat type. According to their criteria, the Ordovician was a time when more disparate trilobite taxa adopted a wider variety of morphotypes than at any other time, before or after (see also Foote 1991). This model is adopted here as the basis for discussion, although it would be surprising indeed if there were no modifications to this scenario in the future; nonetheless, it provides an explicit basis for this summary.

Pelagic Biofacies Ordovician pelagic trilobites, mostly bearing enlarged (“hypertrophied”) eyes, were polyphyletically derived. The pelagic biotope was more richly populated in the Ordovician than in the Late Cambrian and was never reestablished in the Silurian or later, and thus it is a characteristic component of the Ordovician Radiation. The taxa involved belong to the Ibex Fauna and Whiterock Fauna of Adrain et al. (1998). Pelagic trilobites included both epipelagic and mesopelagic species (McCormick and Fortey 1998). The former included some of the most widespread of all trilobites across biofacies; the latter were particularly characteristic of sites marginal to Ordovician paleocontinents, to which they were confined until late in the Ordovician. The mesopelagic community was preserved in the Cyclopygid biofacies and was remarkable for its stability. The earliest record is in the Tremadocian of

Argentina. The early history of the biofacies is entirely peri-Gondwanan. From the Arenig (Fortey and Owens 1987) to the Ashgill (Apollonov 1974) the Cyclopygid biofacies is taxonomically conservative, with the eponymous family dominating the trilobites. Fewer than 10 genera are present, and 6 (Cyclopyge, Microparia, Degamella, Sagavia, Ellipsotaphrus, and a pricyclopygine) are to be found in virtually all these faunas. The fact that rarer, but distinctive, genera such as Gastropolus Whittard also have ArenigAshgill ranges indicates that the whole biotope may eventually be known virtually throughout the Ordovician. These genera are so conservative in morphology that it can be difficult to discriminate an Ashgill from an Arenig species. Cyclopygids are accompanied by the bizarre pelagic bohemillids—now considered aberrant remopleuridioids—and by the enigmatic Cremastoglottos; the latter having as long a stratigraphic range as any cyclopygid. The mesopelagic habitat therefore persisted without a temporal break virtually throughout the Ordovician. The Cyclopygid biofacies did not survive into the Silurian, another line of evidence proving an oceanic crisis at the end of the Ashgill. It had no successor, and so it is absolutely diagnostic of the Ordovician Radiation. Pre-Ashgill Ordovician occurrences were in Avalonia, South America, central Europe, China, and Kazakhstan; in the Ashgill, typical Cylopygid biofacies are known from Girvan, Scotland, and from Quebec, proving that it had crossed the Iapetus remnant into tropical paleolatitudes by then. Symphysops is known from “mound” faunas at that time (Dean 1974), and it is conceivable that the cyclopygids had extended their bathymetric range toward the end of their history. Although very occasional cyclopyids can be found in shallower biofacies in the early Ordovician, they are mostly “stragglers.” The epipelagic biotope is typified by Telephinidae (Carolinites, Oopsites, Opipeuter, Telephina, and Phorocephala). The first three named did not survive the Mid Ordovician (Whiterockian) and were pan-tropical. One species (Carolinites genacinaca) has been described from North America, Siberia, China, and Australia (McCormick and Fortey 1999). Telephina and Phorocephala have early Laurentian records but apparently extended their ranges into high paleolatitudes by the Late Ordovician. No epipelagic type survived into the Silurian.

Trilobites

Olenid Biofacies The olenid biofacies is typified by sulfide-rich, laminated black shale/dark limestone lithologies that accumulated under critically low oxygen conditions and included a restricted fauna adapted to these (Henningsmoen 1957), possibly including specialists capable of living symbiotically with sulfur bacteria (Fortey 2000). The biofacies continues uninterrupted through the Cambrian-Ordovician boundary and is thus composed of Cambrian-style taxa. Olenidae, often with many segments, low convexity, and wide thoraces, constitute the eponymous family. It is controversial whether the agnostoids that may co-occur with them were co-benthic. Other taxa were recruited into the Olenid biofacies through the Ordovician and at the same time assumed olenimorph morphological features, including alsataspidids (Seleneceme), Dionididae (Aethedionide), and remopleuridids (Robergia). Multiplication of segments in some of these forms is spectacular. All are of Ibex Fauna type. It has been recognized that some members of this biofacies (e.g., Bienvillia, Parabolinella, Hypermecaspis) were more or less independent of paleogeography. The Olenid biofacies typically is low gamma diversity/ high individuals of species. The terminal Ordovician crisis that extirpated the Cyclopygidae also eliminated the Olenid biofacies. Deep-sea oxygenated water entrained in oceanic overturn may have been the crucial factor in eliminating the appropriate environment. However, before the end of the Llandovery, olenimorphs (e.g., Aulacopleura) had reappeared, and so this trilobite habitat and biofacies is not uniquely CambroOrdovician.

Filter-Feeding Trilobites Generally small trilobites having a vaulted cephalic chamber flanked by genal prolongations, thorax suspended above the sediment surface, weak axial musculature, “elevated” hypostome, and (usually) reduced eyes have been interpreted as having lived by filtering edible particles from suspension in a feeding chamber. Fortey and Owens (1990a) termed this the “trinucleimorph” design, and indeed this distinctive morphology is exemplified by the Trinucleidae, a diagnostically Ordovician family. However, it did not

first appear in trinucleids, and more than one Cambrian family likely to be only remotely related to Trinucleidae also included genera that showed trinucleimorph design. In the Ordovician, Raphiophoridae, Harpetidae, and Dionididae also typically adopted this morphology. Of these, the first two named are known to be present in the Tremadocian and have plausible Cambrian relatives; trinucleids and dionidids are known from Arenig and younger strata. Rarely, similar morphologies were adopted from other families; for example, the bathyurid Madaraspis was a Laurentian endemic found in strata otherwise lacking species with trinucleimorph design. The early Trinucleidae were overwhelmingly Gondwanan in distribution, achieving more global distribution by the later Llanvirn (late Whiterockian). Raphiophoridae, Harpetidae, and Dionididae seem to be pandemic from the first, and certain genera belonging to these families (e.g., Ampyx, Dionide) are as widely distributed. Among harpetids some genera (Eoharpes) are confined broadly to high paleolatitudes, others (Hibbertia) to low ones. By contrast, trinucleids tend to endemicity. Avalonia, for example, has at least 12 endemic genera (admittedly they may be “oversplit”); other endemic trinucleids are present on the Precordillera terrane of Argentina, which likewise enjoyed an independent history as a microplate. Some raphiophorids are as local: peculiar few-segmented forms such as Taklamakania and the distinctive Bulbaspis are confined to eastern Gondwana and common only in Tarim and Kazakh terranes. Elevated total diversity curves for these trilobites at times of terrane separation will be influenced by the addition of such local taxa. However, high endemicity sits ill with the premise that the asaphoid larvae of trinucleimorphs were planktonic in habit (Chatterton and Speyer 1989), unless this planktonic phase was exceptionally short lived. Suspension-feeding trilobites were not confined to one biofacies but are commonest in outer, or at least quieter, shelf environments, which is to be expected given their substrate preferences. They are typical of the Raphiophorid biofacies around Gondwana (Wales: Fortey and Owens 1978; Argentina: Waisfeld 1995) and Nileid biofacies or equivalent in Laurentia, South China, and Baltica. Although they have reduced eyes or are blind, they may be found associated with other trilobites bearing normal eyes. They may be abundant:

251

252

        .           . Ampyxina is commonly dominant on shaley limestones of Mid Ordovician age in Virginia, and the present author has observed black limestones of slightly younger age largely composed of Taklamakania in northeastern Kazakhstan and Tarim (see also Zhou et al. 1994). The great majority of Ordovician suspension feeders were members of the Ibex Fauna (Adrain et al. 1998; Adrain, Westrop, and Fortey, earlier in this chapter). Only one trinucleoid genus, Raphiophorus, survived the end Ordovician extinction, persisting until the Ludlow. Apart from the long-ranging harpetids, no other candidate for this life habit is known from the Llandovery, but later proetide-derived genera such as Cordania probably adopted it, and so it is not peculiarly Ordovician. However, it is true to say that these kinds of trilobites never again achieved the numerical abundance that they did during the Ordovician biodiversification event.

Particle Feeders Small to middle-sized trilobites with natant hypostomes have been identified as sediment ingesters or particle feeders or both. They are abundant in the Cambrian among the paraphyletic “ptychoparioids.” During that time they may be found in any water depth. In the earlier Ordovician (Ibexian) the habit is represented mostly among Cambrian-style “survivors,” and in Laurentia, Siberia, North China, and Australia it is exemplified by trilobites traditionally assigned to the Hystricurinae. These are mostly found in rather shallow water deposits and may be associated with carbonate muds and silts. Some of the related Dimeropygidae may have had similar habits. In the Tremadocian of Gondwana, in fine-grained clastic deposits for the most part, some olenids that extended beyond the Olenid biofacies, and the eulomatids, were probably the ecologic equivalents of the Laurentian particle feeders. However, after the early Arenig there is a dearth of such forms in Gondwana, the reason for which is unclear. An important change occurs at the base of the Middle Ordovician, when Proetoidea and Aulacopleuroidea (Whiterock Fauna) appear with this morphology and remain almost its sole exemplars until the extinction of the Trilobita. It is clear that these trilobites must have had Cambrian ancestors, but the appearance, radiation, and spread

of these small trilobites through the Mid and Late Ordovician faunas remains a striking phenomenon. They were little affected by the end Ordovician extinction event and are familiar components of Silurian faunas. They also appear to have been tolerant of a variety of biofacies but are most varied and numerous in carbonate “mound” biofacies such as the Boda Limestone (Ashgill) in Sweden.

Predators/Scavengers Trilobites with large, rigidly attached, often buttressed hypostomes with modified posterior borders (forks, burrs, and the like) are attributed to this life mode. There is clearly a variety of specializations within this general category of which we have as yet only speculative ideas. Phacopoids, with specialized visual systems, are considered together with highly distinctive odontopleurids, and it is likely that there were subdivisions with regard to nocturnal or diurnal habits, prey type, and so on, about which we know nothing. However, it is clear that the largest trilobites are of this type. Asaphids, in particular, include very large and robust trilobites, consistent with a position near the top of the food chain, and these are joined by Lichida in the later Ordovician, both groups having pronounced hypostomal forks. The largest trilobites in the Ordovician trilobite faunas of Bohemia (Nobiliasaphus), Avalonia (Basilicus), Iberia (Uralichas), North China (Eoisotelus), and Laurentia (Isotelus) conform to this type. Trilobites of this giant kind are apparently found in inshore habitats, but predator/scavenger morphology can be found in any biofacies except the typical Olenid biofacies. The Early to Mid Ordovician inshore Neseuretus biofacies of Gondwana (and its temporal successors) is dominated by trilobites with conterminant hypostomes having calymenoid, dalmanitoid, and asaphoid affinities. Hammann (1985) suggested that some trilobites with elevated eyes such as homalonotids may have been capable of burrowing. In deeper-shelf biofacies these groups are accompanied by trilobites with other feeding modes, continuing downslope to the atheloptic biofacies, in which the exemplars are often blind or have much reduced eyes. In Early Ordovician Laurentia, cheiruroids (Pliomeridae, Cheiruridae) accompany asaphids in the shallower environments and are joined by caly-

Trilobites menoids, Phacopina, Encrinuridae, Odontopleurida, and Lichida in younger strata. Asaphids are particularly varied in the shelf limestones of the Baltic paleocontinent. Asaphida are part of the Ibex Fauna, which diminished progressively in importance through the Ordovician. However, asaphids (Isotelus) remain conspicuous in shelf limestones in the Cincinnatian of North America. None survived the end Ordovician extinction event, and the proliferation of this group across Ordovician shelf seas can be regarded as a characteristic part of the “great biodiversification.” However, it is striking that the components of the Whiterock Fauna of Adrain et al. (1998) that survived the Ordovician-Silurian event included an array of suborders/families with predator/scavenger morphology that were among the most important components of post-Ordovician faunas: dalmanitids, phacopids, cheirurids, encrinurids, styginids, Lichida, and Odontopleurida among them (figure 24.2). Hence trilobites with attached hypostomes—with the addition of the natant Proetida—survived the Ordovician to provide the basis for subsequent trilobite evolution.

Atheloptic Trilobites Fortey and Owens (1987) coined this term for a “community” of deep-water trilobites that were blind or with eyes much reduced, the majority of which also had close relatives with normal eyes inhabiting more shoreward paleoenvironments. In the Ordovician they were often accompanied by such trilobites as shumardiids and raphiophorids, which lacked eyes in the whole clade. Shumardiids were the last of the miniaturized (at about a millimeter in length) benthic trilobites, which were more diverse in the Cambrian and underwent a modest Ordovician radiation; they may have been particle feeders. They did not survive the Ordovician. The earliest atheloptic assemblage is Tremadocian from northern England (Rushton 1988). Examples are known from all the “series”: Arenig of South Wales, Llanvirn of Bohemia, Caradoc of Kazakhstan, and Ashgill of North Wales. A blind dalmanitoid, Songxites, is known from the Hirnantian of Dob’s Linn, Scotland, and Ireland (Siveter et al. 1980). The atheloptic assemblage of genera typically includes a mixture of Ibex Fauna clades (e.g., nileids such as Illaenopsis, shumardiids) and Whiterock ones

(dindymenines, cheirurines, dalmanitoids). This biofacies, however, was apparently erased at the end of the Ordovician for a period lasting at least through the earlier half of the Silurian. Atheloptic biofacies are well developed again in the Devonian, but none of the proetides or phacopides occupying it then is closely related to the Ordovician examples. ■

Summary ( JMA, GDE, RAF, ØH, JRL, TMcC, AWO, BGW, BDW, SRW, ZYZ)

Cluster analysis of all Ordovician trilobite families using nine biostratigraphic intervals shows that groups followed one of three diversity trajectories. Ibex Fauna I was successful during the Tremadocian and then rapidly declined; Ibex Fauna II radiated during the Arenig and then declined rapidly; the Whiterock Fauna radiated during the Ordovician diversification and was successful through the Ordovician. All the families that survived the end Ordovician mass extinction were members of the Whiterock Fauna. An Ordovician radiation of trilobites occurred globally, except for the tropics of Gondwana, and there was a strong geographic pattern to the radiation. Groups centered in high-latitude Gondwana began to radiate early in the Arenig, and the majority did not survive the end Ordovician. Tropical Laurentian (and Siberian) groups radiated during the late Arenig/early Whiterockian, and almost all survived the extinction. The Whiterock Fauna first dominated craton-margin environments in Laurentia and then spread onshore and offshore during the rest of the Ordovician. Regional diversity curves for Australasia, South America, Avalonia, Baltica, and South China reflect some of the paleogeographic subtleties of the biodiversity change of Ordovician trilobites across a spectrum of tectonic and latitudinal settings. Ordovician trilobites occupied a variety of benthic niches in shallow to deep-water habitats and colonized the open seas. The pelagic habitat was remarkably persistent but did not survive the end Ordovician extinction event. The Olenid biofacies and atheloptic habitat were similarly affected, but their ecologic equivalents are known (with unrelated taxa) from younger strata. Among shelf faunas, the Ordovician was characterized by a proliferation of suspensionfeeding genera; proetides were the principal deposit feeders from the Whiterockian onward. Members of

253

254

        .           . the Ibex Fauna dominated the pelagic and Olenid biofacies and included also some of the largest predatory/ scavenger taxa. The majority of trilobites of the Whiterock Fauna that survived the end Ordovician extinction had attached hypostomes.  We are grateful to Bob Owens and John Shergold for their helpful and supportive comments on an ear-

lier version of this chapter. J. M. Adrain and S. R. Westrop’s research is supported by NSF grant EAR 9973065. A. W. Owen and T. McCormick’s work on the Anglo-Welsh trilobites was funded by NERC Grant GR3/11834, which is gratefully acknowledged. B. G. Waisfeld acknowledges support from CONICET, ANPCyT, and Fundación Antorchas. Z.-Y. Zhou’s research on Chinese trilobites was financially supported by the Major State Basic Research Development Program (No. G2000077700).

25

Eurypterids, Phyllocarids, and Ostracodes Simon J. Braddy,Victor P. Tollerton Jr., Patrick R. Racheboeuf, and Roger Schallreuter

O

verviews of three arthropod groups, the eurypterids, phyllocarids, and ostracodes, are included here. Ordovician eurypterids are rare. Reliable records, mostly from North America, include megalograptids and rarer carcinosomatids, stylonurids, and erieopterids. Of the 31 species described from the Ordovician of New York State, only one is a eurypterid, one is a phyllocarid, and the remainder comprise indistinguishable remains (pseudofossils). Eurypterid taxonomy and phylogeny are in a state of flux; calculations of diversity measures are thus inappropriate. Ordovician phyllocarid crustaceans are widely distributed in graptolitic black shale facies, but documentation of the group remains inadequate. Biodiversity studies are limited to a preliminary survey of the phyllocarids from the Ordovician of South America presented herein. Ostracodes are small crustaceans whose body is enclosed within a carapace consisting of two valves normally of calcareous composition, except for planktonic species with mainly organic shells. The body is strongly compressed laterally and also reduced in length to consist of head and thorax only; the abdomen is missing or only rudimentary, and the number of appendages is reduced to seven, the smallest number among all crustaceans. Extant ostracodes inhabit nearly all aquatic environments. Ordovician representatives appear to be confined to marine settings (although this may be limited by lack of recog-

nition of lacustrine habitats of the time); to date, true pelagic forms have not been recognized. ■

Eurypterids (SJB, VPT)

The order Eurypterida forms a small ( blastoids and bothriocidaroids > echinoids; see figures 26.3 and 26.4). Finally, a few echinoderm groups, including diploporans and cornute stylophorans, had unusual diversity

275

276

    .  patterns that do not closely resemble those of most other groups. Diploporans (figure 26.3) were the most diverse group of blastozoans and ranged throughout almost the entire Ordovician, but they had an earlier peak in diversity in the Mid Ordovician TS.4c than other groups and had somewhat less diversity through the Mohawkian, when the diversity of other blastozoan and crinozoan groups peaked. Paul (1988) pointed out that diploporans were a heterogeneous and very likely polyphyletic grouping of blastozoans with similar but convergent pore structures. Consequently, they should probably be broken up into several monophyletic subgroups. Because many diploporans were stemless and either recumbent in soft muddy sediment or directly cemented to a hard substrate (Parsley 1988), they had a somewhat different life mode than most other stemmed blastozoans. Also possibly contributing to their different diversity pattern is that diploporans were apparently most common in coolwater (temperate) clastic environments. Cornute stylophorans (figure 26.4) are a Cambrian group that continued into the Ordovician, branched to the more advanced and bilaterally symmetrical mitrate stylophorans apparently in TS.1b, and then reached their peak diversity in Early Ordovician TS.2b. This peak diversity is based almost entirely on the diverse cornutes in the St.-Chinian Formation in the Montagne Noire region of southern France (Vizcaïno and Lefebvre 1999). Although common earlier, cornutes had already disappeared from Laurentia by this time. As mitrate stylophorans slowly expanded during the Mid Ordovician, cornute stylophorans gradually declined, becoming rare in the Mohawkian (mid Caradoc), when most other echinoderm groups reached their peak diversity. They made their last appearance in the early and mid Ashgill before becoming extinct. This distinctive Ordovician diversity pattern for cornute stylophorans appears to mark a relict group that was replaced by more advanced mitrate stylophorans. ■

New Designs and Life Modes

Several divergent echinoderm body plans and new life modes appeared abruptly at various times during the Ordovician. The echinoderm groups having these new body plans are characterized by numerous morphological innovations as compared with their Cam-

brian predecessors. Few morphological intermediates link these new groups to possible ancestors, although some early members (called precursors by Sprinkle 1995) did not have the fully formed morphology characteristic of most later members. Somasteroids (Spencer and Wright 1966) were a small precursor group to brittle stars and starfish (Blake and Guensburg 1993). These asterozoans were the first free-living echinoderms with a star-shaped body made up only of axial and perforate extraxial skeletal components (Mooi and David 2000). They had a downward-facing mouth, enlarged food grooves, and biserial ambulacral and adambulacral plates. All three asterozoan groups appeared sequentially in the Early Ordovician, and a Late Ordovician asteroid has been found with its arms wrapped in feeding position around a smaller bivalve, indicating a carnivorous life mode like that of many modern asteroids (Blake and Guensburg 1994). Another new echinoderm design is shown by ophiocistioids that appeared in the Mid Ordovician, followed by bothriocidarids and echinoids in the early Late Ordovician. Ophiocistioids had a globular to biscuit-shaped test with partly organized plate columns, medium to long ambulacra bearing large pores for tube feet(?), a downward-facing mouth containing a lantern with teeth for rasping, and a lateral to upward-facing anus. In the two echinoidlike clades, the tube foot pores are smaller and usually paired, and articulated spines are mounted on pustules (later on bosses) on the columns of interambulacral and ambulacral test plates. The times of origin and possible phylogenetic relationships of ophiocistioides, bothriocidarids, echinoids, and sclerite-bearing holothurians have been discussed by Smith (1988b) and Smith and Savill (2001). The first three clades appear to have been mobile, epifaunal herbivores and scavengers; the earliest holothurians may have been epifaunal deposit feeders. Surviving CEF blastozoan eocrinoids and newly appearing PEF crinoids and blastozoan groups such as glyptocystitid rhombiferans, hemicosmitid rhombiferans, and parablastoids developed similar body plans as stalked or stemmed, medium- to high-level, upright suspension feeders. All these clades standardized the thecal plating by reducing the numerous irregular plates and growing the remaining thecal plates larger (Sprinkle and Guensburg 2001). These

Echinoderms larger thecal plates formed a few alternating circlets, developed better pentameral symmetry, and provided for stronger articulation at the stem facet (basals or infrabasals), better support of erect arms or recumbent ambulacra (radials), and protection of the central mouth (orals). Many Late Cambrian echinoderms respired through thin, nearly smooth thecal plates (Sprinkle 1973a; Sumrall et al. 1997). Most new PEF groups in the Early and Mid Ordovician changed this design to thicker and more ridged thecal plates that had specialized respiratory pores or folds shared between adjacent plates (many blastozoans, few crinoids). Ambulacral plating in blastozoans was standardized to an arrangement where one to two biserial floor plates supported each brachiole, and early crinoids developed erect arms having both an axial component and extraxial-derived brachials (Guensburg and Sprinkle 2001). These new crinoid arms converged on the “erect ambulacra” of Late Cambrian trachelocrinid eocrinoids, Mid to Late Ordovician hemicosmitid rhombiferans, and Late Ordovician coronoids (Bockelie 1979; Brett et al. 1983; Sumrall et al. 1997). However, crinoid arms were better organized and more efficient, especially when most camerate crinoids added closely spaced pinnules to the arms starting in the Mid Ordovician, soon followed by some disparids and cladids. The tegmens of the earliest crinoids were similar to the oral surfaces of ancestral edrioasteroids (Guensburg and Sprinkle 2001), but some camerate crinoids soon developed sunken ambulacra and thick cover plates that resembled adjacent tegmen plates, thus hiding the tegmen food grooves and providing more protection against predators or parasites. The tiny-plated stalks of some early crinoids (Guensburg 1992) were initially organized into pentameres (earliest Ordovician) and later into one-piece columnals (latest Early Ordovician), convergent with those developed earlier in the Mid Cambrian by blastozoan eocrinoids (Sprinkle 1973a). Holdfasts at the distal tips of stems were differentiated for attachment to both hard and soft substrates (Brett 1981) or discarded in adults so the stem could be wrapped around other objects (camerate crinoids; Guensburg 1992) or to allow the adult echinoderm to become free living or recumbent on the seafloor (many glyptocystitid rhombiferans, some rhipidocystids, and a few crinoids; Paul 1984; Lewis et al. 1987).

■

Echinoderm Provincial Patterns

Another finding of the project, evident in the monographs and faunal lists that were used to assemble the database, is that the composition of the richest echinoderm faunas varied considerably from one occurrence to another. Figure 26.5 is a series of histograms plotting the number of species present in different echinoderm clades for five of the richest Ordovician echinoderm faunas surveyed for the project. All of these range from 40 to 79 echinoderm species, with the most diverse faunas in the Barrandian of Bohemia (figure 26.5C) spanning nearly the entire Ordovician (TS.1a–6c). The Fillmore and Wah Wah formations in western Laurentia (figure 26.5A) span most of the Early Ordovician Ibexian (TS.1b–2c), and these units have a balanced and moderately diverse fauna with about 15 percent each of disparids, glyptocystitid rhombiferans, and edrioasteroids but only one homalozoan (a mitrate) (Guensburg and Sprinkle 1994; Sprinkle and Guensburg 1995, 1997). The Montagne Noire region of southern France (figure 26.5B), on the northwestern edge of Gondwana in high to intermediate latitudes (chapter 5), has six formations spanning the Tremadocian and Arenig (TS.1a–4a), the lower half of the Ordovician, and this region has produced diverse echinoderm faunas dominated by cornute (41 percent) and to a lesser degree mitrate (16 percent) stylophorans but only two crinoids (Vizcaïno and Lefebvre 1999). The Barrandian fauna of Bohemia (figure 26.5C), a periGondwanan terrane in high latitudes (chapter 5), has 12 formations spanning nearly the entire Ordovician and bearing several diverse echinoderm faunas dominated by diploporans (24 percent), mitrate stylophorans (17 percent), and edrioasteroids (11 percent) (Prokop and Petr 1999). The Lebanon Formation in Tennessee (central Laurentia) (figure 26.5D) has a moderately diverse early Mohawkian (early Late Ordovician; TS.5b) echinoderm fauna dominated by crinoids, especially diplobathrid camerates (30 percent) and disparids (16 percent), but lacking homalozoans (Guensburg 1984). Finally, the Lady Burn Starfish Bed in western Scotland (figure 26.5E), which was then an island belt perhaps adjacent to northeastern Laurentia, has produced a large and varied echinoderm fauna from a few thin sandstone beds in the Late Ordovician mid Ashgill (TS.6a) that has

277

Number of Species

10

A

Filmore–Wah Wah, western United States, TS.1b–2c

B

Montagne Noire, southern France, TS.1a–4a

C

Bohemia–Barrandian, TS.1a–6c

D

Lebanon Limestone, central United States, TS.5b

E

Lady Burn Starfish Bed, Girvan, Scotland, TS.6b

5 0

20 15 10 5 0

Number of Species

20 15 10 5 0

15 10 5 0 10

numerous ophiuroids (18 percent) and asteroids (13 percent), plus bothriocidarids and most of the echinoids (together 13 percent) known from the Ordovician (Donovan et al. 1996; Jefferies and Daley 1996). Different echinoderm clades dominated these Ordovician faunas depending on their age, geographic location, climate zonation, and facies and environmental setting. Several authors have discussed the paleogeography, climatic zonation, and migration routes of Ordovician echinoderms including Paul (1976) for all echinoderms, Eckert (1988) for American and British crinoids, and Donovan (1989) for British crinoids. Paul (1976) identified three faunal provinces for Ordovician echinoderms, based on earlier work using trilobites and brachiopods: (1) North American (now commonly called Laurentian); (2) Baltic, including Scandinavia, Estonia, and western Russia; and (3) South European (now commonly called Northwestern Gondwanan), including Morocco, Spain, southern France, southern Britain, Bohemia, and perhaps some localities farther east in the

5

Cornute Stylophorans Mitrate Stylophorans

Edrioasteroids Cyclocystoids Somasteroids Asteroids Ophiuroids Bothriocidarids Echinoids

Homoiosteleans Eocrinoids Glyptocystitid Rhombiferans Caryocystitid Rhombiferans Diploporans Parablastoids Rhipidocystids Paracrinoids Coronoids

0 Stem Crinoids Diplobathrid Camerates Monobathrid Camerates Disparids Cladids Hybocrinids Flexibles

Number of Species

FIGURE 26.5. Composition of five relatively diverse echinoderm faunas from different geographic areas and parts of the Ordovician. A, Ibexian echinoderms from the Fillmore and Wah Wah formations in the western United States; note the fairly even distribution of groups except for only one homalozoan (Sprinkle and Guensburg 1997). B, Tremadocian and Arenig echinoderms from the Montagne Noire region of southern France; note the dominance of cornute and mitrate stylophorans but only two crinoids (Vizcaïno and Lefebvre 1999). C, Early to Late Ordovician echinoderms from the Barrandian region of Bohemia (Czech Republic); note the dominance by diploporans and mitrate stylophorans and fairly even distribution of other echinoderms except for relatively few crinoids (Prokop and Petr 1999). D, Mohawkian echinoderms from the Lebanon Limestone in the central United States; note the dominance by diplobathrid camerates, relatively few blastozoans, and no homalozoans (Guensburg 1984). E, Middle Ashgill echinoderms from the Lady Burn Starfish Bed of Scotland; note the large numbers of ophiuroids, asteroids, bothriocidarids, and echinoids and fairly even distribution of other groups except for relatively few blastozoans (Donovan et al. 1996; Jefferies and Daley 1996).

Number of Species

    . 

Number of Species

278

“Tethyan” region. The Laurentian Province was tropical (within 0–30 degrees of the equator), the Baltic Province drifted from intermediate latitudes in the Early Ordovician to low latitudes in the Late Ordovician, and the Northwestern Gondwanan Province was temperate to polar (at 40–80 degrees south) (Paul 1976), although large ice caps were present on nearby land only in the latest Ordovician (Hirnantian). During the Early Ordovician, these three faunal provinces were well differentiated and showed little mixing of echinoderms, many of which were endemic. In the Mid Ordovician, the provinces were still recognizable but showed more faunal migrations and a greater mixture of echinoderms. By the Late Ordovician, the faunal provinces were less distinct, and there were many cosmopolitan echinoderm families and genera (Paul 1976). This pattern probably represents both the geographic spread of genera and species as they diversified during the Ordovician and the gradual closure of part of the Iapetus Ocean by plate movements, bringing the faunal components of the Lau-

Echinoderms rentian Province and at least one peri-Gondwanan terrane into contact with Northwestern Gondwanan provincal elements by the Late Ordovician. In the Early and Mid Ordovician, the Laurentian Province had numerous crinoids, glyptocystitid rhombiferans, eocrinoids, paracrinoids, parablastoids, and edrioasteroids but relatively few or no diploporans, hemicosmitids, caryocystitids, coronoids, homoiosteleans, and cornute and mitrate stylophorans (figure 26.5A, D). In contrast, the Northwestern Gondwanan Province had numerous diploporans, glyptocystitid rhombiferans, caryocystitids, eocrinoids, edrioasteroids, somasteroids, and cornute and mitrate stylophorans but relatively few or no crinoids, paracrinoids, parablastoids, and homoiosteleans (figure 26.5B, C). By the latest Ordovician, the fauna from Scotland included a mixture of echinoderms from several faunal provinces, such as crinoids, glyptocystitid rhombiferans, asteroids, and ophiuroids, plus newly evolved groups such as bothriocidarids and echinoids and a few relict clades such as cornute stylophorans, minus relict groups that had already become extinct or were very rare, such as rhipidocystids, parablastoids, and paracrinoids (figure 26.5E). ■

Conclusions

1. Echinoderms diversified greatly during the Ordovician as new and more advanced members of the PEF added to and replaced older members of the CEF. Echinoderm diversity increased from 8–9 classes in the CEF to a peak of 17 classes and 29 distinctive clades in the PEF by the early Late Ordovician. 2. The echinoderm component changed from a small eocrinoid and stylophoran-dominated fauna during the Late Cambrian to much larger crinoid, rhombiferan, and diploporan-dominated faunas during the Ordovician. Crinoids then dominated the echinoderm fossil record from the Late Ordovician to the end of the Paleozoic. 3. In terms of the percentage of completeness of the echinoderm fossil record in the Ordovician, the 30 clades plotted over the 19 time slices had an average of 32 percent gaps (implying 68 percent occurrence). Most long-lived echinoderm clades had relatively complete occurrence records, but some closely related shorter-lived clades had very divergent occurrence patterns.

4. The echinoderm radiation began near the beginning of the Early Ordovician, during which 21 of the 30 echinoderm clades appeared, and continued until the early Late Ordovician, when all the echinoderm classes and Paleozoic clades had appeared. This Early Ordovician appearance of echinoderms preceded that of many other metazoan groups in the PEF, although the time of peak diversity was similar. 5. There was little change in species diversity in most echinoderm clades at the Early Ordovician–Mid Ordovician boundary, when many other metazoan groups diversified, and at the Mid Ordovician–Late Ordovician boundary. This implies that many echinoderm clades only slowly increased in diversity during the Early and Mid Ordovician. 6. However, there was a dramatic diversity increase in crinoids and many other echinoderm clades in the early Late Ordovician (Mohawkian or mid Caradoc), when echinoderms reached their peak species and clade diversity for the Ordovician and had developed nearly all their Paleozoic life modes. Although six new clades and several new life modes appeared at this time, they did not contribute much to overall diversity. Instead, there was a much larger diversity increase of clades (especially in crinoids) that had appeared much earlier. 7. Diversity then declined during the rest of the Late Ordovician as smaller and less advanced PEF echinoderm classes dropped out of the record. Echinoderms were scarce in the latest Ordovician (Hirnantian), even though 23 of 27 clades present in the Ashgill survived this glacially driven extinction interval. 8. Advanced echinoderm clades show a sequential appearance, asteroids during the Early Ordovician, echinoids and holothurians during the Late Ordovician, perhaps indicating their likely time of branching and rapid morphological change from ancestral stem group to intermediate precursor to standardized advanced clade. 9. A few groups show unusual diversity patterns, including diploporans, which have an earlier diversity peak than most other echinoderms, and cornute stylophorans, which also peaked early and were then replaced by more advanced mitrate stylophorans before becoming extinct in the Late Ordovician. Both diploporans and cornutes were most common in deeper-water clastics in temperate regions where other echinoderms were less common.

279

280

    .  10. The composition of the richest echinoderm faunas varied considerably from one occurrence to another, based on time of deposition during the Ordovician, geographic location of the echinoderm fauna, type of facies and depositional setting, and climatic zonation.

 We thank Sergei V. Rozhnov (Paleontological Institute, Moscow, Russia), David L. Meyer (University of Cincinnati, Cincinnati, Ohio), Colin D. Sumrall (University of Tennessee, Knoxville, Tennessee), and Mark McKinzie (Grapevine, Texas) for assistance in compiling the echinoderm diversity of particular geographic areas for which they had detailed knowledge. Our part of the chapter is based on field work

supported by the National Science Foundation under Grants BSR-8906568 (Sprinkle) and EAR-9304253 (Sprinkle, Guensburg), by CRDF Cooperative Research Grant RG1-242 (Rozhnov, Sprinkle, Guensburg), and by a Petroleum Research Fund, American Chemical Society Grant to Mark Wilson, College of Wooster (Guensburg). We also thank the University of Texas Geology Foundation, Austin, Texas, and Rock Valley College, Rockford, Illinois, for additional funds for travel and manuscript preparation. Chris Schneider, University of Texas at Austin, prepared the diversity plots in figures 26.1–26.5 from the spreadsheets and reviewed an early draft of the manuscript. Barry Webby (Macquarie University, Sydney, Australia), Peter Jell (Queensland Museum, Brisbane, Australia), and a third, anonymous reviewer read the completed manuscript and offered many helpful suggestions.

27

Graptolites: Patterns of Diversity Across Paleolatitudes Roger A. Cooper, Jörg Maletz, Lindsey Taylor, and Jan A. Zalasiewicz

G

raptolites (Graptoloidea or planktic Graptolithina) provide an ideal group for the study of biodiversity through the Ordovician because they were widely distributed around the globe and are well represented in numerous, relatively continuous black shale sequences. They originated at the base of the Ordovician and became nearly extinct at the top, thus providing a closed system and removing the problem of “edge effects” (Foote 2000a). In addition, because they are widely used for dating and correlation, the stratigraphic ranges of species are generally well known. The level of taxonomic and biostratigraphic investigation internationally ranges widely in quality. We use the complete species lists for Australasia and Avalonia, each of which regions have complete or relatively complete graptolite zonal suites through the Ordovician. In the Early to Middle Ordovician Avalonia occupied a relatively high paleolatitude (Scotese and McKerrow 1990; chapter 5), whereas Australasia lay in low paleolatitudes throughout the Ordovician. For comparison with a region that was, during the Early to Middle Ordovician, in middle paleolatitudes, the less complete graptolite sequence of Baltica is also used. In this study we attempt to distinguish between those features of diversity and evolutionary rates that were affected by latitude and those that were global. In a parallel study (Sadler and Cooper unpubl.) a global survey of Or-

dovician and Silurian graptolite successions is analyzed using the constrained optimization (CONOP) method to derive the global pattern of graptolite species diversity change. ■

Data Sets

The species lists and zonal range charts for the three regions have been compiled, updated, and revised by the authors. Species qualified by a query (?) in the source faunal lists are omitted. Species listed with a “cf.” or “aff.” are included only where it is clear that they are not synonymous with other species in the list for the region. Subspecies are omitted unless they represent a distinct stratigraphic horizon or are as morphologically distinctive as an average species. Thus the four subspecies of Isograptus manubriatus in the Australasian Yapeenian Ya1 and Ya2 zones (figure 27.1) are entered as a single taxon. Dichograptus maccoyi maccoyi (Bendigonian Be1 Zone) and D. m. densus (Chewtonian Ch1 Zone) are separate taxa. However, the five successional subspecies of Isograptus victoriae are combined into three taxa by grouping adjacent pairs. This is because Cooper (1973) found, in a multivariate analysis, that adjacent subspecies overlapped by up to 40 percent. The overlap reduced to zero or nearly zero (i.e., comparable with discrete species) when alternate subspecies were omitted. The

281

AVALONIA (SOUTHERN BRITAIN)

INTERNATIONAL

6b 6a

D.anceps

6

N. persculptus/extraordinarius

ASHGILL

6c

P. pacificus

5b

CARADOC

5

Pleurograptus linearis

5a

D.clingani

LATE

5d

No graptolites

D.complexus

Dicellogr. complanatus

5c

BALTICA

Dicellograptus morrisi Ensigraptus caudatus

AUSTRALASIA

Time units

 .   .

Bo4-5 N.persc./extraor.

21

Bo3

P.pacificus

20

Bo2

(pre-pacificus) C.uncinatus

Bo1 Ea4 Ea3 Pleurograptus linearis upper Ea2 D.clingani lower Ea1

Mesograptus multidens

Mesograptus multidens

Nemagraptus gracilis

Nemagraptus gracilis

D.gravis D.kirki D.spiniferus D.lanceolatus

AGE (Ma) 443

19 18 17

450

16

Gi2 O.calcaratus

15

Gi1

N.gracilis

14

Da4

P.riddellensis

13

Aulograptus cucullus (hirundo)

Undulogr. austrodentatus

Da1 U.austrodent.

3b

Isograptus gibberulus

Cardiograptus Oncograptus

Ya1-2 Ca3-4

3a

2b 2a

1

1c 1b 1a

12 11 10

Oncograptus I.v.max-maximo.

9

Ca1-2 I.v.lunatus-vict.

8

Ch1-2

D.protobifidus

7

Didymograptus varicosus Expansogr. protobalticus

Be1-4

P.fruticosus

6

Tetragr. phyllograptoides Tetragr. phyllograptoides

La3

Isograptus victoriae Didymograptus simulans

Arenigr.hastatus - A. gracilis

I. v. lunatus - I.v. victoriae Baltograpus minutus

T.approximatus

Araneograptus murrayi

Araneograptus murrayi B. ramosus K. kiaeri

(sedgwickii & salopiensis) Adelograptus "tenellus" Rhabdinopora f.anglica Rhabdinopora f.flabelliformis Rhabdinopora f.parabola

Adelograptus

sp.

(R.f.anglica) A.matanensis R.f.parabola R.praeparabola

470

5

Hunnegraptus copiosus

1d TREMADOC

EARLY

artus

Pterograptus elegans

upper Nicholsonogr. fasciculatus Da3 P.decoratus middle Holmograptus lentus lower Da2 U.intersitus Undulogr. dentatus

4a

2c

2

Didymograptus

H. teretiusculus/P. distichus

La2

3

4b

Hustedo. teretiusculus Didymograptus murchisoni

A.victoriae

4

LLANVIRN

4c

ARENIG

DARRIWILIAN

MIDDLE

460

TREMADOCIAN

282

480 (A.pulchellum)

(P.antiquus)

La1b

Psigraptus

La1a

Anisograptus

(pre-La)

4

3 2 1

489

FIGURE 27.1. Correlation of Ordovician graptolite zones (sources given in text) and time units used in this study. Alternate units are shaded. Age calibration of Australasian zones from Sadler and Cooper (chapter 3). Dashed lines in Britain and Baltica columns = correlation with Australasia uncertain. Dashed line in Australasian column = time calibration of boundary uncertain. Short-ranging zones are combined, and long ones split, in order to achieve time units of more even duration. The time units used here, numbered at right, are preferred to time slices recommended by Webby et al. (chapter 2, listed at left) because their boundaries are more readily located in the graptolite zonal succession.

taxa used here are therefore all loosely of “speciesequivalent” rank and are referred to as “species.” In all three regions the stratigraphic ranges of several species within the time units used here are known, making it possible to refine the age range of taxa. It also enables some regional zones to be split to accord with the time units adopted here. The first and last appearances of species in the regional data sets should be regarded as regional events rather than global events. As noted by Cooper et al. (1991), an equivalent of the deep-water graptolite biofacies of Australasia and

other low paleolatitude regions, representing slope and base of slope depositional environments, is lacking from Avalonia and other intermediate to high paleolatitude regions, also from Baltica. It is unclear at present whether this lack is because the facies simply was not developed in these regions or whether it was likely to have been developed but is not now preserved. If the latter, our estimate of diversity in these regions is likely to be too low. Responsibility for the regional data sets is as follows: Australasia, R. A. Cooper and A. H. M. Vanden-

Graptolites Berg; Baltica, J. Maletz; Avalonia, J. A. Zalasiewicz and L. Taylor.

Australasia The Australasian region, comprising Australia and New Zealand, includes the classic graptolitic sections described by Hall, Harris, Thomas, Benson, and Keble (see VandenBerg and Cooper 1992 for references), particularly in the Lachlan Fold Belt of Australia and its equivalent in New Zealand. The sedimentary succession comprises turbidites, cherts, and shales throughout the Ordovician, and the black shale facies is developed in all zones except for the Bolindian Bo5 Zone. The earliest Tremadocian international graptolite chronozones, R. praeparabola and R. f. parabola, however, are not represented by graptolites, and there are likely to be faunal gaps at the zonal or subzonal level higher in the Tremadocian (Cooper 1999a; Maletz and Egenhoff 2001). Diversity through the early Tremadocian is therefore possibly underestimated. The sequence of zones, particularly the postTremadocian zones, has become a standard of reference within the Pacific faunal realm. The graptolite sequence has been fully reviewed by VandenBerg and Cooper, who list 30 zones, 2 of which are subdivided into subzones. The zones are grouped here into 26 time units. The very long Lancefieldian La2 Zone, A. victoriae, is informally split into upper and lower parts, using published (Cooper 1979; Cooper and Stewart 1979) and unpublished information on the stratigraphic ranges of species within the zone, in order to better match with the time unit divisions. VandenBerg and Cooper (1992) listed 313 named species and subspecies, reduced here to 283 taxa of “species-equivalent” rank. Graptolitic shales are well developed in all time units (not to be confused with the 19 global time slices also employed in this volume; see figure 27.1).

Baltica A total of 213 taxa is recorded for Baltica, which is here taken to include Sweden, Denmark, Estonia, and Norway and to exclude the richly graptolitic Bogo Shale of the Trondheim region, which was not part of the Baltica paleoplate (Dewey et al. 1970; Schmidt-

Gündel 1994). Graptolites of the northern German boreholes in the Rügen area are also omitted, as it is not certain to which paleoplate these belong. A suite of 31 zones is used for plotting species ranges (Maletz 1995, 1997, unpubl.), and these are grouped here into 18 time units (figure 27.1). One zone, D. clingani, is informally split into upper and lower parts, using knowledge of the range of species within the zone. A problem in sampling the graptolite fauna of Baltica for diversity studies is the limited number of stratigraphic sections and the unfavorable facies in some parts of the column. The graptolites in some stratigraphic intervals have been sampled and studied in detail, for example, the “Dictyonema Shales” and “Lower” and “Upper Didymograptus Shales” of earlier authors. On the other hand, other parts of the column, such as the Upper Ordovician zones, are little studied. The regrouping of zones into longer time units reduces the effects of unevenness in sampling intensity and biostratigraphic study to some extent. However, in time units 9, 10, and 11 (the Cardiograptus, U. austrodentatus, and U. dentatus zones of the late Arenig) the graptolite facies in most of the OsloScania shale belt is replaced by a carbonate facies, the Komstad Limestone Formation and its equivalents, that is less favorable for preservation of graptolites. This interval is yet to be fully investigated for graptolites, and its apparently depressed graptolite diversity may therefore be largely an artifact of nonpreservation or of sampling deficiency. For these reasons, we give the Baltica data set less weight than the other two sets in interpretation. On the other hand, the preservation quality of Baltic graptolites is exceptionally high, and the Tremadocian succession of the Oslo region is, in addition, among the stratigraphically best controlled in the world. Stratigraphic ranges of graptolites in the classic and richly diverse Töyen Shale (formerly known as the Lower Didymograptus Shale) used here are those revised by Maletz (unpubl.). Graptolites of late Caradocian and Ashgillian age are not known in Scandinavia, and time units 22 to 26 are therefore unrepresented.

Avalonia This region includes England and Wales and contains the classic graptolite-bearing strata studied by

283

284

 .   . Lapworth, Elles, and Wood. Two hundred and twelve species are included in the present analysis. The graptolite census of Strachan (1996, 1997) has been revised and updated in the light of new work and the zonal ranges replotted in terms of the currently used zonal scheme (figure 27.1). The zonal scheme largely follows that of Fortey et al. (2000). In the middle Tremadocian (salopiensis and sedgwickii trilobite zones), equivalent to the B. ramosus and K. kiaeri zones of Baltica, a widespread facies change and unconformity greatly reduce graptolite field occurrence, affecting the diversity pattern. Similarly, no graptolites are recorded in the late Caradoc equivalent of the D. complanatus Zone of Scotland and of the Ea4 Zone, D. gravis, of Australasia. Otherwise, the graptolitic facies is developed within (but not necessarily throughout) all time units in Avalonia. The D. artus Zone is split informally into two parts in order to align better with the time unit divisions used here (figure 27.1).

Reliability and Comprehensiveness Of the three regions, Australasia and Britain have graptolite localities that number in the thousands, and they have numerous sections spanning successions of zones. Although exact numbers are unknown, it is unlikely that Baltica contains a similar abundance of collecting localities throughout the Ordovician. In this respect Australasia and Britain are likely to be more representative than Baltica of the original graptolite faunas that lived in the regions. This may be the reason for the mean duration of species in Baltica being appreciably shorter than in both Australasia and Britain (see later in this chapter). A relatively smaller number of collecting localities makes it possible that sampling and recorded stratigraphic ranges are less complete than in the other two regions. An artificially shortened mean species duration would affect diversity and evolutionary rate patterns. ■

Ordovician Timescale

A reliable and precise timescale for the Ordovician is necessary for the accurate representation of evolutionary rates. Graptolite stratigraphic ranges are expressed in zones in each region and the zonal schemes

are correlated in figure 27.1. The zones must be calibrated in millions of years to derive rates. The graphic correlation-based composite timescale of Cooper (1999b) was an attempt to do this. It was based on 10 relatively long ranging, Early to Mid Ordovician, high-quality, deep-water graptolitic measured sections, representing two paleoplates, and on the conodontbased graphic correlation of 61 Late Ordovician, carbonate sections in North America (Sweet 1984, 1988a). A much larger database, comprising 1,136 taxa in almost 200 deep-water shale measured sections from around the world, has been used in a novel application of the CONOP procedure by Sadler and Cooper (unpubl.) to derive a composite sequence that stands as a proxy relative timescale for the Ordovician and Silurian periods. In the biostratigraphic database are graptolite assemblages that are reliably tied to 22 highresolution zircon dates. These are used, first, to test the linearity of the relative timescale and, second, to calibrate the scale. The method is outlined by Sadler and Cooper (chapter 3 and unpubl.). It is the scale used in this chapter. Of the six new international time divisions for the Ordovician (figure 27.1) only the Darriwilian and the Tremadocian have been formally named. The Caradoc and Ashgill can be used informally, with their British definitions (Fortey et al. 1995), for the Late Ordovician. The gap between the end of the Tremadocian and start of the Darriwilian is here informally referred to as Arenig, so that we can refer to a complete set of global Ordovician divisions.

Time Units Ideally, evolutionary rates and diversity trends should be measured against a scale with time units of equal duration to avoid the distorting effects of uneven duration (Sepkoski and Koch 1996). This is generally not practical for most paleontological data, which are recorded in zones or stages that are of unknown or uncertain relative duration or are of known but uneven duration. Although the relative duration of zones in each of the three regions studied here is known with unusually good precision, the zones range widely in duration (figure 27.1). Therefore, zones of short duration (such as the four zones of the Australasian Bendigonian) are grouped into a single time

Graptolites unit, and long zones (such as the La2 Zone of Australasia and D. artus Zone of Avalonia) are split in two. The correlation of the three regional graptolitic zonal successions (figure 27.1) follows Webby et al. (chapter 2). Time unit boundaries are made to coincide with zone boundaries (or subzone boundaries), and therefore some time units vary slightly from region to region. The result is a scale of 21 time units that, in duration, are more even than zones in any of the zonal schemes, thereby minimizing any distortion of measured rates. Time units average 2.195 million years (m.y.) in duration (σ = 0.89 m.y.) and range in duration from 1.3 m.y. to 5.1 m.y.

Mean Species Duration Using the CONOP timescale (chapter 3), the mean duration of a species in each of the three regions, with standard deviation and total number of species recorded from the region, is as follows: Australasia Baltica Avalonia ■

2.382 m.y. 1.43 m.y. 2.377 m.y.

σ = 1.894 σ = 1.09 σ = 1.821

283 total 210 total 212 total

Measures of Diversity and Taxonomic Rates Possible Sources of Bias

The present species lists have been compiled by the authors and their colleagues from published and unpublished sources and checked for taxonomic consistency, and the stratigraphic ranges of species revised in the light of the current zonal schemes. Although inconsistencies and errors are certain to remain in our data, we believe that they are unlikely to affect significantly the patterns emerging from the analysis, at least for Avalonia and Australasia. As mentioned already, the patterns for Baltica are interpreted with more caution. Other sources of bias arise from uneven outcrop area and sampling opportunity and uneven or incomplete preservation quality and collecting completeness. As discussed earlier, these are believed to be minimal in our data sets, especially those of Avalonia and Australasia. The possible effects of inaccurate or uneven duration of time units are discussed in the next section.

Effects of Time Unit Duration Uncertainty or pronounced unevenness (or both) in time unit duration can have a distorting effect on the pattern of diversity change through time (Sepkoski and Koch 1996; Foote 2000a). Sepkoski and Koch recommend minimizing this problem by combining zones of short duration and subdividing zones of long duration in order to even out the differences in time unit durations. With a relatively precise timescale and a well-constrained zonal correlation of sequences, we are able to plot the stratigraphic ranges of species for the three regions in a common scale of time units that are of relatively uniform duration, as discussed earlier. We therefore believe that the distorting effects of uncertain and uneven time units are minimized in the present study. Consistent with this view, we note that in the total diversity and total faunal turnover curves (see the next section), the main peaks do not coincide with the longest time units; nor do the main troughs coincide with the shortest time units. Furthermore, by using artificial data sets, Cooper (chapter 4) was able to directly test and compare the performance of three alternative measures of diversity on time units of uneven duration and thereby compare them with “true” mean standing diversity (MSD). Foote (2000a) has modeled the effects on measures of diversity, and origination and extinction rates, of interval length, preservation quality, “edge effects,” and the taxonomic rates themselves. He concludes that single-interval taxa produce many undesirable distortions and recommends that diversity and rate measures be adopted that do not count singleinterval taxa (taxa confined to a single time unit). However, to ignore these taxa in our data sets would be to ignore a large proportion of the database. Some of the concerns raised by Foote may be minimized in the present study. As mentioned earlier, “edge effects” are largely removed, and we are using a precise timescale with subequal time units, thereby reducing the distorting effects of time interval durations. Variable preservation quality and collection completeness will undoubtedly affect the data. Incomplete collecting will tend to shorten true ranges and reduce diversity, as is possibly the case in Baltica. It will also be unlikely to detect the rare species, affecting diversity as well as origination and extinction rates. The zonal

285

286

 .   . ranges of taxa, especially in Avalonia and Australasia, however, are well “filled in” (there are few gaps throughout the range; VandenBerg and Cooper 1992; Taylor and Zalasiewicz unpubl.), testifying to the completeness of collecting. In any case, it is the trend of the diversity curve, rather than the absolute value of the diversity measure, that is important here.

Appropriate Diversity Measures Cooper (chapter 4) constructed six model data sets in which the durations of time units and of species life spans are comparable with those found in our graptolite data sets. He then measured diversity in the data sets utilizing three commonly used estimates of diversity: total diversity, species per m.y., and normalized diversity. Total diversity (dtot ) is the total number of species that are recorded from the time unit; it is the simplest and most commonly used measure for estimating MSD. Species per m.y. (di ) is simply total diversity divided by the duration of the time unit (i ) in m.y.; it allows for the probability that longer time units will capture more species. Normalized diversity (dnorm ) is a measure that normalizes for variability in time unit duration to the extent that the longer a time unit is, the more species will begin or end in it or are confined to it. It is the sum of species that range from the unit below to the unit above, plus half the number of species that range beyond the time unit but originate or become extinct within it, plus half those that are confined to the time unit itself (Sepkoski 1975). The measure also compensates for another bias. Species ranges in regional data sets are generally expressed as zonal ranges. These will overestimate true ranges because few species ranges will completely span the zones in which they first appear or last appear or to which they are confined. Because the “true” time ranges of species in the data sets are known, the “true” MSD of successive time units can be calculated. This was then compared with each of the three estimates of diversity in turn. The results show that the total diversity (dtot ) measure consistently and substantially overestimates true MSD, deviating from it by 25 percent on average, and the species per m.y. (di ) measure generally underestimates it, deviating by 29 percent on average. The

normalized diversity (dnorm ) measure is consistently the closest approximation of true MSD, deviating by 10 percent on average. In our regional analyses we give the three measures of diversity just described (dtot , di , and dnorm ), but, unless otherwise stated, comparisons and conclusions are based on the normalized measure (dnorm ), the best estimator of MSD.

Origination and Extinction Measures Origination and extinction intensity and rate measures are included in this study of diversity because, apart from the effects of regional migration, diversity change is a function of the dynamics of origination and extinction. Three measures of extinction are given: the total number of extinction events in the time unit; the percentage of extinction, which is the number of extinction events divided by the total number of taxa present; and the per capita extinction rate, which is the percentage of extinction divided by the duration of the time unit, in m.y. (Sepkoski and Koch 1996). The equivalent measures for origination are also given: Number of originations Percentage of origination Per capita rate of origination Number of extinctions Percentage of extinction Per capita rate of extinction

=o = od × 100 = odi =e = ed × 100 = edi

Faunal Turnover This is the faunal turnover that takes place in a time unit. Three measures are given. The first, total faunal turnover, is expressed as the sum of the number of extinctions and originations in a time unit. The second is the proportion or percentage of faunal turnover; it is total faunal turnover divided by twice the total diversity for the time unit. The third is the per capita rate of faunal turnover, which is the percentage of faunal turnover divided by the duration of the time unit, in m.y. Faunal turnover = (o + e) Percentage of faunal turnover = (o + e)2d × 100 Per capita rate of faunal turnover = (o + e)2di (in m.y.)

Graptolites ■

Diversity Patterns

Diversity patterns for the three regions are shown in figures 27.2, 27.3, and 27.4, and the basic statistics for diversity and evolutionary change are given in table 27.1. Features common to all three regions are likely to reflect global patterns, whereas contrasts between Avalonia and Australasia may reflect latitudinal effects. Although the graptolite clade appeared with a burst at the base of the Tremadocian, rapidly invaded a range of biotopes in the oceans, and spread widely around the globe (Bulman 1971; Cooper 1999a), it was not rich in species for a long time. Graptolites originated as planktic organisms at the beginning of Ordovician time. The earliest species are found in facies representing continental slope water depths, and within 2 m.y. the clade had expanded its habitat range into farther offshore (and possibly deeper)

biotopes, as well as into the epipelagic zone, in which it ranges from inner shelf to outer slope (Cooper 1998, 1999a). Yet diversity was slow to expand. For the first 12 m.y. (25 percent) of Ordovician time, normalized mean standing species diversity remains relatively low, at no more than 9 species. Then, within the space of 3 m.y. (time unit 6 in the early Arenig), there is a dramatic expansion, as if some threshold had been crossed, and diversity reaches a maximum or near maximum for the entire Ordovician. This is most marked in Australasia, where there is a dramatic expansion in the Bendigonian stage, and normalized diversity reaches 37.5 species (total species diversity reaches 73). The sudden expansion is also seen in the species per m.y. curve (figure 27.2A). It is clearly not an effect of time unit duration. In Avalonia, the effect is less strong, but there is a rapid expansion in the early Arenig, and normalized species diversity reaches 22.3 (total diversity 33). The low

A

B 60

total ( d tot)

60 50

Species

Species

70

normalized ( d norm)

40 30 10

40

e

30

10

species per m.y. (d i)

0

0

D

C Per Capita Rate

Faunal turnover (e+o)

100

Species

o

50

20

20

80 60 40 20

Origination

percentage (od)

1.0

100%

rate (odi)

0.8

80

0.6

60

0.4

40

0.2

20

0.0

0

F

E Faunal turnover

1.0

percentage (o+e )2d

0.8

100%

rate (o+e)2di

80

0.6

60

0.4

40

0.2

20

0.0

1

3

TREMADOC

5

7

9

ARENIG

11

13

DARRIWIL.

15

17

19

21

CARADOC ASHGILL

Extinction

Per Capita Rate

Per Capita Rate

Origination, extinction (e, o)

Diversity

1.0 0.8

100%

rate (edi)

percentage (ed)

80 60

0.6

40

0.4

20

0.2 0.0

1

3

5

TREMADOC

Time Units

FIGURE 27.2. Diversity and rate plots for Australasia. There are no data for time unit 1.

7

9

ARENIG

11

13 15 17 19

DARRIWIL.

Time Units

21

CARADOC ASHGILL

287

 .   . A

B

50

Diversity

45

normalized (d norm)

35

total ( d )

25 20 15

20

e

15 5

sp./m.y.(di)

5

1 C

3

5

7

9

0

11 13 15 17

Faunal turnover ( o+e )

1 D

5

7

9

Origination

3

5

7

9

80

0.6

60

0.4

0

11 13 15 17

40

rate (odi)

1

3

5

7

20

9 11 13 15 17

F Extinction

Faunal turnover

percentage (o + e)2d

0.8

80

0.6

60

0.4

40

rate (o+ e)2di

0.2

20

percentage ( ed)

1.0

100%

Per Capita

1.0

100%

0.8

0.2

1

11 13 15 17

percentage ( od)

1

E

0.0

3

75

Per Capita

Species

25

10

10

60 50 40 30 20 10 0

o

30

30

0

Originations, extinctions

35

Species

Species

40

Per Capita

288

100%

0.8

80

0.6

60

0.4

40

rate (e e di)

0.2

20

0.0

1

3

TREMADOCIAN

5

7 ARENIG

9

11 13 15 17 DARRIWILIAN

CARADOC

Time Units

1

3

TREMADOCIAN

5

7

9

ARENIG

11 13 15 17 DARRIWILIAN

CARADOC

Time Units

FIGURE 27.3. Diversity and rate plots for Baltica. There is little, or no, information in time units 9–11 inclusive.

diversity throughout the Tremadocian and rapid increase in the early Arenig are features common to all regions. They are likely to be global and not significantly influenced by latitude. The more extreme expansion in diversity in Australasia, however, may have been a latitudinal effect. On the other hand, the reduction in diversity during the Late Ordovician, which led eventually to the near extinction of the group in the late Ashgill, appears to have begun much earlier in Avalonia, and possibly Baltica, than in Australasia. In Avalonia, a steep decline begins in the mid Caradoc, and diversity never recovers before the end of the Ordovician. In Australasia, the decline does not begin until the mid Ashgill. During the mid Caradoc to Ashgill, therefore, MSD in Avalonia is significantly lower than in Australasia (figure 27.5). This contributes to the lower overall total Ordovician diversity in Avalonia (212 species) being significantly lower than in Australasia (289 species). If latitude is the cause of

this difference, it presumably operates either directly through surface water temperature (Skevington 1974) or indirectly through the earlier breakdown of ocean density structure in high paleolatitudes associated with global cooling, leading to decay of high productivity zones such as developed in the oxygen minimum zone and along continental margins due to upwelling, both regarded as favorable habitats for graptolites (Berry et al. 1987; Finney and Berry 1997; Cooper 1998). However, according to the plate tectonic model of Cocks and Torsvik (chapter 5), the northward drift of Baltica and Avalonia carried both regions into lower latitudes by Late Ordovician time. Although they lay in opposite hemispheres, Australasia and Avalonia would have lain a similar distance from the Ordovician equator by the late Caradoc. In terms of this model, latitude may have had little effect, and the reason for the earlier decline in diversity in Avalonia remains uncertain.

70 60 50 40 30 20 10 0

A

B Originations, extinctions

Diversity 30

total (d )

normalized ( d norm)

sp./m.y.

Species

Species

Graptolites

(d i)

20

0

D Faunal turnover (e + o)

50

Origination

1.0

Per Capita

Species

o

10

C

40 30 20 10 0

percentage (od)

0.8

rate (odi)

0.6

100% 80 60

0.4

40

0.2

20

0.0

E

F Faunal turnover

100%

percentage (o+ e)

2d

0.8

Per Capita

1.0

Per Capita

e

80

0.6

60

rate (o+ e)

0.4

2di

40

0.2

20

0.0

Extinction

1.0

percentage ( ed) rate (edi)

0.8

100% 80

0.6

60

0.4

40

0.2

20

0.0

1

3

TREMADOC

5

7

9

ARENIG

11

13

DARRIWIL.

15

17 19

21

1

3

TREMADOC

CARADOC ASHGILL

Time Units

5

7

9

ARENIG

11 13 15 17 19 21 DARRIWIL. CARADOC ASHGILL

Time Units

FIGURE 27.4. Diversity and rate plots for Avalonia. There are no data in time unit 4 and the D. complanatus Zone (time unit 18).

A comparison of the normalized diversity curves for Australasia and Avalonia (figure 27.5) reveals that the commonly held view, that standing diversity through the Ordovician was significantly lower in high paleolatitudes (Skevington 1974; Berry 1979; Cooper et al. 1991), is only partly true. In the early to mid Arenig, when latitudinal difference was most marked, the difference in diversity is indeed marked, being appreciably higher in Australasia (figure 27.5). From the mid Caradoc time on, the paleolatitudinal difference between the two regions may have been insignificant, and the observed difference between them in normalized diversity could not, therefore, be related to latitude. For most of the remainder of the Ordovician, normalized diversity is as high as, or higher than, that in Australasia. This is all the more significant in view of the lack, in Avalonia, of the deepwater graptolite biofacies (Cooper et al. 1991). The lower overall diversity of Baltica possibly reflects its smaller sampling base and therefore may not be significant.

■

Evolutionary Rates

The curves for numbers of originations (o) and for numbers of extinctions (e) in Australasia (figure 27.2B) fluctuate markedly and are strongly positively correlated (r2 = 0.74). As a result, originations and extinctions act in concert to drive the total faunal turnover curve, which fluctuates strongly. Faunal turnover seldom drops below 20 species in any time unit and reaches more than 100 species in the early Arenig (time unit 6), the highest level recorded globally. The high level of faunal change is what has enabled the fine subdivision of the Australasian graptolite succession into 32 zones and subzones (VandenBerg and Cooper 1992). The same trends can be seen in Baltica (figure 27.3), where the total faunal change curve has a similarly strong fluctuation. In Avalonia (figure 27.4), the origination (o) and extinction (e) curves also fluctuate strongly. They are positively correlated except for where diversity is high. In the late Arenig to mid Caradoc they are negatively correlated

289

2

Note: Time units are defined in Figure 27.1.

2 1.43 1.00 2 1.00 0.00 0.71 2 1.00 0.00 0.71 4 1.00 0.00 0.71

AVALONIA Total diversity (d ) Species/m.y. (di ) Normalized diversity (dnorm ) Originations (o) Proportion of originations (od ) Std error of od Per capita rate of origination (odi ) Extinctions (e) Proportion of extinctions (ed ) Std error of ed Per capita rate of extinction (edi ) Faunal turnover (o + e) Proportion of turnover (o + e)2d Std error of (o + e)2d Per capita rate of turnover (o + e)2di 8 5.00 4.00 8 1.00 0.00 0.63 8 1.00 0.00 0.63 16 1.00 0.00 0.63

8.00 4.65 3.99 7.00 0.88 0.12 0.51 8.00 1.00 0.00 0.58 15.00 0.94 0.06 0.55

3.00 2.73 1.50 3.00 1.00 0.00 0.91 3.00 1.00 0.00 0.91 6.00 1.00 0.00 0.91

BALTICA Total diversity (d ) Species/m.y. (di ) Normalized diversity (dnorm ) Originations (o) Proportion of originations (od ) Std error of od Per capita rate of origination (odi ) Extinctions (e) Proportion of extinctions (ed ) Std error of ed Per capita rate of extinction (edi ) Faunal turnover (o + e) Proportion of turnover (o + e)2d Std error of (o + e)2d Per capita rate of turnover (o + e)2di

488.0 1.7 8.00 1.90 4.00 8 1.00 0.00 0.58 8 1.00 0.00 0.58 16 1.00 0.00 0.58

Age of Base Duration

AUSTRALASIA Total diversity (d ) Species/m.y. (di ) Normalized diversity (dnorm ) Originations (o) Proportion of originations (od ) Std error of od Per capita rate of origination (odi ) Extinctions (e) Proportion of extinctions (ed ) Std error of ed Per capita rate of extinction (edi ) Faunal turnover (o + e) Proportion of turnover (o + e)2d Std error of (o + e)2d Per capita rate of turnover (o + e)2di

1 489.3 0.8

Time Unit

3

1 0.42 0.50 1 1.00 0.00 0.42 1 1.00 0.00 0.42 2 1.00 0.00 0.42

8.00 3.36 4.00 8.00 1.00 0.00 0.42 7.00 0.88 0.12 0.37 15.00 0.94 0.06 0.39

4.00 1.68 2.00 4 1.00 0.00 0.42 0 0.00 0.00 0.00 4 0.50 0.18 0.21

486.3 2.4

Tremadocian 4

3 0.59 1.50 3 1.00 0.00 0.20 3 1.00 0.00 0.20 6 1.00 0.00 0.20

11.00 2.16 5.50 10.00 0.91 0.09 0.18 7.00 0.64 0.15 0.12 17.00 0.77 0.09 0.15

16.00 3.14 9.50 12 0.75 0.11 0.15 8 0.50 0.13 0.10 20 0.63 0.09 0.12

483.9 5.1

5

5 2.08 2.50 5 1.00 0.00 0.42 3 0.60 0.22 0.25 8 0.80 0.13 0.33

18.00 7.50 9.00 14.00 0.78 0.10 0.32 16.00 0.89 0.07 0.37 30.00 0.83 0.06 0.35

15.00 6.30 8.50 6 0.40 0.13 0.17 8 0.53 0.13 0.22 14 0.47 0.09 0.20

478.8 2.4

6

15 5.36 8.50 13 0.87 0.09 0.31 4 0.27 0.11 0.10 17 0.57 0.09 0.20

35.00 12.50 18.00 33.00 0.94 0.04 0.34 22.00 0.63 0.08 0.22 55.00 0.79 0.05 0.28

73.00 25.44 37.50 66 0.90 0.03 0.32 44 0.60 0.06 0.21 110 0.75 0.04 0.26

476.4 2.9

33 12.69 20.00 22 0.67 0.08 0.26 12 0.36 0.08 0.14 34 0.52 0.06 0.20

23.00 12.11 13.50 10.00 0.43 0.10 0.23 14.00 0.61 0.10 0.32 24.00 0.52 0.07 0.27

48.00 25.95 30.00 20 0.42 0.07 0.23 26 0.54 0.07 0.29 46 0.48 0.05 0.26

473.6 1.9

7

Arenig 8

30 15.00 22.33 9 0.30 0.08 0.15 8 0.27 0.08 0.13 17 0.28 0.06 0.14

32.00 16.00 16.00 23.00 0.72 0.08 0.36 28.00 0.88 0.06 0.44 51.00 0.80 0.05 0.40

29.00 14.50 20.50 8 0.28 0.08 0.14 10 0.34 0.09 0.17 18 0.31 0.06 0.16

471.7 2.0

9

35 26.92 21.50 13 0.37 0.08 0.29 24 0.69 0.08 0.53 37 0.53 0.06 0.41

8.00 4.00 4.50 4.00 0.50 0.18 0.25 7.00 0.88 0.12 0.44 11.00 0.69 0.12 0.34

40.00 22.22 23.50 24 0.60 0.08 0.33 22 0.55 0.08 0.31 46 0.58 0.06 0.32

469.7 1.8

10

26 16.25 13.50 15 0.58 0.10 0.36 15 0.58 0.10 0.36 30 0.58 0.07 0.36

5.00 3.13 2.50 4.00 0.80 0.18 0.50 5.00 1.00 0.00 0.63 9.00 0.90 0.09 0.56

25.00 13.89 18.00 8 0.32 0.09 0.18 7 0.28 0.09 0.16 15 0.30 0.06 0.17

467.9 1.8

41 45.56 24.50 30 0.73 0.07 0.81 3 0.07 0.04 0.08 33 0.40 0.05 0.45

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

29.00 21.97 19.00 11 0.38 0.09 0.29 13 0.45 0.09 0.34 24 0.41 0.06 0.31

466.1 1.3

11

40 15.38 22.82 2 0.05 0.03 0.02 31 0.78 0.07 0.30 33 0.41 0.06 0.16

27.00 13.50 13.50 27.00 1.00 0.00 0.50 15.00 0.56 0.10 0.28 42.00 0.78 0.06 0.39

36.00 16.51 19.50 19 0.53 0.08 0.24 30 0.83 0.06 0.38 49 0.68 0.05 0.31

464.8 2.2

12

Darriwilian

33 15.00 18.00 26 0.79 0.07 0.36 15 0.45 0.09 0.21 41 0.62 0.06 0.28

49.00 22.27 25.00 38.00 0.78 0.06 0.35 37.00 0.76 0.06 0.34 75.00 0.77 0.04 0.35

19.00 8.64 10.50 12 0.63 0.11 0.29 9 0.47 0.11 0.22 21 0.55 0.08 0.25

462.6 2.2

13

42 13.55 23.00 24 0.57 0.08 0.18 16 0.38 0.07 0.12 40 0.48 0.05 0.15

23.00 7.42 14.00 11.00 0.48 0.10 0.15 8.00 0.35 0.10 0.11 19.00 0.41 0.07 0.13

21.00 6.77 12.50 11 0.52 0.11 0.17 8 0.38 0.11 0.12 19 0.45 0.08 0.15

460.4 3.1

14

TABLE 27.1. Graptolite Diversity, Origination, Extinction, and Faunal Turnover for Three Data Sets, Australasia, Baltica, and Avalonia

15

43 17.92 23.50 17 0.40 0.07 0.16 31 0.72 0.07 0.30 48 0.56 0.05 0.23

16.00 6.67 9.00 2.00 0.13 0.08 0.05 14.00 0.88 0.08 0.36 16.00 0.50 0.09 0.21

23.00 9.58 13.00 10 0.43 0.10 0.18 13 0.57 0.10 0.24 23 0.50 0.07 0.21

457.3 2.4

24 9.23 13.00 12 0.50 0.10 0.19 20 0.83 0.08 0.32 32 0.67 0.07 0.26

10.00 3.85 6.00 8.00 0.80 0.13 0.31 9.00 0.90 0.09 0.35 17.00 0.85 0.08 0.33

34.00 12.88 18.00 24 0.71 0.08 0.27 24 0.71 0.08 0.27 48 0.71 0.06 0.27

454.9 2.6

16

Caradoc 17

7 5.38 5.64 1 0.14 0.13 0.11 7 1.00 0.00 0.77 8 0.57 0.13 0.44

9.00 6.92 4.50 8.00 0.89 0.10 0.68 9.00 1.00 0.00 0.77 17.00 0.94 0.05 0.73

25.00 19.84 17.00 14 0.56 0.10 0.44 5 0.20 0.08 0.16 19 0.38 0.07 0.30

452.3 1.3

18

0 0.00 0.00 0 0.00 0.00 0.00 0 0.00 0.00 0.00 0 0.00 0.00 0.00

24.00 18.46 16.50 4 0.17 0.08 0.13 12 0.50 0.10 0.38 16 0.33 0.07 0.26

451.0 1.3

19

3 1.00 3.00 3 1.00 0.00 0.33 1 0.33 0.27 0.11 4 0.67 0.19 0.22

25.00 8.33 14.50 13 0.52 0.10 0.17 14 0.56 0.10 0.19 27 0.54 0.07 0.18

449.7 3.0

3 1.43 3.00 1 0.33 0.27 0.16 3 1.00 0.00 0.48 4 0.67 0.19 0.32

14.00 7.37 7.50 3 0.21 0.11 0.11 13 0.93 0.07 0.49 16 0.57 0.09 0.30

446.7 1.9

20

Ashgill 21

7 3.68 8.21 7 1.00 0.00 0.53 7 1.00 0.00 0.53 14 1.00 0.00 0.53

5.00 2.38 3.00 4 0.80 0.18 0.38 5 1.00 0.00 0.48 9 0.90 0.09 0.43

444.8 2.1

International Divisions

Time units

Graptolites

Avalonia Australasia diversity

Avalonia orig

ext

Australasia orig

21

6

Ashgill

Evolutionary Events

ext Hirnantian crash, near elimination of graptolite clade

Hirnantian glaciation, 5 Late Ashgill regression, 2,3,5

Extinction rate peak

Late Caradoc regressive event, 2,3

Highstand SL, 2,5

19 17

5 Caradoc 15

Expansion and dominance of "diplograptids," waning of diversity in high latitudes

13

4 Darriwilian

11 9

3 Arenig

7

2 5

1 Tremad-

Extinction of most dichograptids Short-lived burst of dichograptids, sinograptids Appearance of “diplograptids” Diversification of the isograptids, followed by extinction of isograptids and dichograptids

1

Carbon 13 peak, 4 Highstand SL, 2,3 Transgression, 2,3 Late Darriwilian regression, 2,3 Highstand SL, 3 Late Arenig regression, 1,2,3

Highstand SL, 1,2,3 Evolutionary expansion of graptolites, driven by proliferation of Major global transgression, dichograptids, sigmagraptids 1,2

3

ocian

Environmental Events

Expansion of habitat range, low diversity Appearance of planktic graptolites

Several sharp regressive events, 1 Carbon 13 peak, 6

FIGURE 27.5. Comparison of normalized diversity plots, and per capita rates of origination (orig) and extinction (ext), for Australasia and Avalonia, showing main graptolite macroevolutionary events and events affecting the marine environment. Key to references: 1, Nielsen (1992a); 2, Ross and Ross (1995); 3, Fortey (1984); 4, Ainsaar et al. (1999); 5, Brenchley et al. (1994); 6, Ripperdan and Miller (1995).

(r 2 = 0.31). The faunal turnover curve is much smoother than in the other two regions as a result. Faunal turnover is high (>30 species) from late Arenig to early Caradoc, reaching a high of 45 species in the mid Caradoc (Mesograptus multidens Zone). When the numbers of extinctions and originations are normalized for total diversity, a similar but less marked pattern is seen in the Australasian data (figures 27.2D, F). The origination percentage (od ) and extinction percentage (ed ) curves fluctuate strongly, driving a strongly fluctuating faunal turnover percentage curve (o + e)2d . Ignoring the “end effects” in the earliest and latest Ordovician resulting from low total diversity, faunal turnover (figure 27.2E) averages about 50 percent of the species in a time unit and peaks at about 70 percent. Although overall correlation between origination and extinction is now very weak (r 2 = 0.1), there is an accordance of peaks in time units 6 (early Arenig), 9 (late Arenig), and 16 (mid Caradoc). In Baltica, the pattern is less clear, but faunal turnover percentage remains high throughout. However, in both Baltica and Avalonia, the marked

accordance of peaks and troughs seen in the Australasian data set is lacking, and the faunal turnover percentage curves are smoother. Fluctuations in origination and extinction numbers are less strong in Australasia when normalized for both diversity and time unit duration. These curves are the per capita rates of origination (odi ), extinction (edi ), and faunal turnover ((o + e)2di ), shown in figure 27.2D–F. They suggest that the stronger fluctuations in the percentage and total numbers curves may, at least in part, be due to the effects of variation in time unit duration. Most conspicuously, the big peak in evolutionary activity in the early Arenig (time unit 6, Bendigonian stage), coinciding with the sharp rise in normalized diversity (figure 27.2A), is all but invisible. This suggests that although there are many extinctions and originations in this time unit, when expressed as a percentage of the total number present and normalized for duration of the time unit, it is nothing remarkable. However, the Bendigonian stage contains a high proportion of species with short stratigraphic ranges, spanning no more than one or

291

 .   . two of the four zones (VandenBerg and Cooper 1992). We suspect that when the pattern is analyzed at a scale finer than the time units used here, it may change. In Baltica, the per capita rates of origination and extinction are highest in the late Arenig–early Darriwilian (figure 27.3D–F), but they lie to either side of a sampling gap in the data. The per capita rate of faunal turnover peaks at 0.56, in the early Darriwilian (time unit 10). In Avalonia, both the percentage and per capita rate curves for originations and extinctions fluctuate strongly. The per capita rate of origination peaks in the lower D. artus Zone (Darriwilian) at 0.8, the highest level reached at any time in any region. Similarly, the per capita rate of extinction peaks in the lower P. linearis Zone (late Caradoc) at 0.8, higher than in any other region. In time units where diversity is high (mid Arenig to mid Caradoc), there is weak negative correlation between the percentage rates of origination and extinction (r 2 = 0.34), and the per capita rates of origination and extinction are uncorrelated (r 2 = 0.01). Numbers of originations and extinctions are, of course, highly dependent on total diversity. Their decline in the earliest and latest Ordovician in all regions is due to the dropoff in total diversity at these times. However, the strong fluctuations in rates through the Ordovician in the two evolutionary processes are seen even after allowing for variation in total diversity and time unit duration and are likely to be significant. If the negative correlation of origination and extinction observed in the mid Arenig to mid Caradoc in Avalonia is significant, it contrasts with the other regions where positive correlation dominates. However, it is unlikely to be related to latitudinal differences, which diminished during this time. Most noticeable is the dissimilarity between Australasia and Avalonia when either their extinction rate or their origination rate curves are compared (figure 27.5, table 27.2). Although the per capita rates of origination in the two regions are similar in having strong fluctuations, they are only weakly correlated (r 2 = 0.24). The per capita rates of extinction in the two regions also fluctuate strongly and are uncorrelated (r 2 = 0.07). Apart from the latest Arenig (time unit 9) and mid Darriwilian (time unit 12), where extinction peaks are present in both regions, there is little obvious accordance in either curves in the two

TABLE 27.2. Correlation of Origination and Extinction Rates (r 2 values) Per Capita Rate of Origination Per Capita Rate of Extinction

292

Australasia Baltica Avalonia

Australasia

Baltica

Avalonia

0.01 0.07

0.24 0.72

0.24 0.02 -

regions. Unless there is an undetected bias in our data, this dissimilarity must reflect dissimilar origination and extinction patterns in the two regions. Because the dissimilarity persists through the Ordovician, it is unlikely to be related to latitude. When rate of origination in Baltica is compared with rates of origination in the other two regions (table 27.2), there is similarly no strong correlation. Extinction rate in Baltica is uncorrelated with that in Australasia but, interestingly, is strongly correlated with that in Avalonia (r 2 = 0.72), with which region it is most closely geographically located. If the correlation coefficients can be taken at face value, the pattern suggests that origination rates are not related to either latitude or geographic proximity. Extinction rates are not strongly related to latitude but are related to geographic proximity, as might be expected. If the rates are influenced by environmental parameters such as nutrient supply and eutrophication (Allmon and Ross 2001), then these parameters vary, at least to a moderate extent, independently in Australasia and Avalonia. ■

Environmental Events

The normalized diversity curves and per capita rate curves for origination and extinction for Avalonia and Australasia are compared in figure 27.5. The main macroevolutionary events are compared in time with some major events that affect the ocean environment. The most spectacular evolutionary event (common to all three regions) is the rapid diversification in the early Arenig, marked by highs in the origination percentage curves for each region and with a diversity maximum for the entire Ordovician in Australasia. This coincides with a global transgression and highstand sea level (Nielsen 1992a), conditions that would suit the widespread development of the zone in the water column of oxygen depletion

Graptolites and nitrogen enrichment, recognized as a favored habitat for graptolites (Berry et al. 1987; Cooper et al. 1991; Cooper 1998) along with warm saline bottom waters in a greenhouse world (Wilde and Berry 1986). The expansion was driven by a proliferation of the Dichograptidae and Sigmagraptidae, following the decline of the Anisograptidae. Highstand sea level has also been inferred in the mid Darriwilian (Fortey 1984), mid Caradoc (Ross and Ross 1995), and early Ashgill (Brenchley et al. 1994), all times with a high in the percentage origination curves. In the mid Darriwilian, a diverse dichograptacean and glossograptacean fauna, including didymograptids, sigmagraptids, sinograptids, isograptids, and glossograptids, makes a short-lived appearance. In the mid Caradoc, many diplograptids and orthograptids appear, and corynoidids become locally abundant. Ironically, the high productivity and oceanographic conditions associated with at least one of these highstands (mid Caradoc) is likely to have produced high extinction rates among the marine benthos on the continental shelves (Patzkowsky et al. 1997). On the other hand, regressive events, such as those recognized in the late Arenig, late Caradoc, and possibly late Darriwilian, are matched by prominent extinction events, although not simultaneously in the three regions. The late Arenig marks the extinction of many isograptids and dichograptids. In the late Caradoc, many species of diplograptids, orthograptids, and Dicranograptidae became extinct. The major regression in the late Ashgill, associated with the Hirnantian glaciation (Brenchley et al. 1994), brought about the near extinction of the entire graptolite clade (Koren 1991; Melchin and Mitchell 1991). Only three or four lineages survived to give rise to Silurian stock. During times of regression associated with climatic minima, density stratification and structure of the oceans break down, ocean circulation and turnover increase, the oceans become ventilated, and the nitrogen-rich, oxygen minimum zone decays (Wilde and Berry 1984). This would reduce the extent of the habitat zone for graptolites, consistent with the ob-

served extinction pattern (Melchin and Mitchell 1991). Its effect on the distribution and extent of marginal upwelling zones, favored by Finney and Berry (1997) as the preferred habitat for graptolites, has yet to be assessed, however. ■

Conclusions

1. The mean duration of a species in Avalonia (2.377 m.y.) is not significantly different from that in Australasia (2.382 m.y.). As compared with Australasia, MSD in Avalonia is lower in the early to mid Arenig and late Caradoc–Ashgill and higher for much of the remainder of the Ordovician. 2. Origination and extinction rates fluctuate strongly through the Ordovician in all three regions. In Avalonia they appear to be negatively correlated during the time of high diversity (mid Arenig to mid Caradoc), whereas in the other regions positive correlation prevails. 3. Many of the major macroevolutionary extinction, origination, and diversity change events are matched by oceanic environmental events; transgressions and highstand sea levels are generally times of increased origination, and regressions are times of increased extinction. These effects are likely to be global. 4. The only feature detected that is possibly related to latitude is the rate of increase in diversity in the Early Ordovician; it is more rapid and reaches a higher level in Australasia (low latitude) than in Avalonia (high latitude). In terms of the plate tectonic model of Cocks and Torsvik (chapter 5), differences between the two regions in the Late Ordovician, such as the earlier decline in diversity and negative correlation of origination and extinction rates in Avalonia, are unlikely to be related to latitude.

 We thank Dan Goldman and Fons VandenBerg for their reviews of the manuscript. This project was supported in part by the Foundation for Research, Science and Technology, in New Zealand.

293

28

Chitinozoans Florentin Paris, Aïcha Achab, Esther Asselin, Chen Xiao-hong, Yngve Grahn, Jaak Nõlvak, Olga Obut, Joakim Samuelsson, Nikolai Sennikov, Marco Vecoli, Jacques Verniers, Wang Xiao-feng, and Theresa Winchester-Seeto

C

hitinozoans are an extinct group of organicwalled microfossils. The earliest known species appeared in the Early Ordovician (Tremadocian), and the group became extinct at the end of the Devonian. Chitinozoans have been reported from most types of Ordovician marine sediments. Major constraints affecting their occurrence and preservation are high-energy hydrodynamic regimes, weathering, and medium- to high-grade metamorphism. Because of their minimal dependence on lithology, chitinozoans are usually continuously present in most Ordovician marine successions, and the first-appearance datum (FAD) and last-appearance datum (LAD) of species can be precisely controlled. Paris and Nõlvak (1999) postulated that chitinozoans are the eggs of soft-bodied marine metazoans and consequently can be used to document the biodiversification of their unknown producers. Chitinozoans diversified rapidly throughout the Ordovician Period. Of the 56 accepted chitinozoan genera (Paris et al. 1999), 63 percent first appeared in the Ordovician (28 percent in the Early Ordovician, 24 percent in the Mid Ordovician, and 11 percent in the Late Ordovician). Only 2 genera became extinct at the end of the Early Ordovician, 5 during the Mid Ordovician, and a total of 19 at the end of the Ordovician. The Ordovician biodiversification is also clearly reflected by the introduction of the major morphological innovations: 65 percent of the most important

294

ones appeared before the end of the Mid Ordovician and 77 percent before the end of the Ordovician (Paris et al. 1999). Chitinozoan generic and specific diversity is low. It normally ranges from a few to rarely more than 10 species representing only a few genera per sample. The abundance of chitinozoans, however, ranges from several hundred to a few thousand specimens per gram of rock for regions occupying high latitudes (e.g., North Gondwana). In general there are no more than a few hundred specimens per gram of rock in regions of low latitude (e.g., Laurentia, East Gondwana, Baltica); however, lithological and other environmental factors also influence the distribution of chitinozoans. Despite their inferred pelagic mode of distribution, chitinozoans display a definite provincialism during the Ordovician (Paris 1981; Achab 1991), when the major continental plates were far apart and the latitudinal contrast was most significant. Individual biodiversity curves have been established for all major paleocontinents where sufficiently abundant and reliable data were available. ■

Data and Counting Methods

Ordovician chitinozoans have been reported from all continents (figure 28.1), except Antarctica (see references in Miller 1996; Paris 1996). By the end of 2000, the global literature devoted to Ordovician

Chitinozoans Greenland Siberia

Arctic W Canada W. Canada

Sibe ria

ia rent

Lau

E. Canada 746

E. & MW. USA: 589

s

tu pe

Russia NW 370

Sweden 1682

W. USA

Baltica

Norway 270

Estonia 2000

Lithuania 350

Ia

ia n o al v A U.K

Altai

Belarus

Tarim

China

Latvia 500

Poland

Korea

Ukraine

S.China 230

Germany

Belgium

c

Rhei

PortuFrance gal 285

Czech .

Iran

Austria Turkey

Portugal

M. East Spain

Mauritania.

Algeria 1105 Niger

Libya

Saudi Arabia

na

Morocco

Gondwa

India S. Africa

Chad

Florida

Australia

South Pole Brazil Bolivia

Argentina

> 500 > 500

250 – - 500

100 – - 250

25 – - 100

< 25

FIGURE 28.1. Regional distribution and density of the Ordovician samples used for documenting the chitinozoan diversity (the main references can be found in Achab 1989; Nõlvak and Grahn 1993; Paris 1996, 1999; Nõlvak 1999; Samuelsson and Verniers 1999; Wang and Chen 1999; Servais and Paris 2000). The paleogeographic reconstruction is based on faunal and sedimentologic evidence. It corresponds to Mid to Late Ordovician time (about 450–455 Ma). Data compiled and map drawn by F. Paris 2002.

chitinozoans included 380 publications (excluding abstracts and theses or other unpublished reports) and contained descriptions of some 397 Ordovician species (including subspecies and forms raised to the specific level). It is possible the Ordovician chitinozoan diversity may reach a total near 500 species. Regional databases have been built for each major paleoplate, using the normalized time slicing proposed by Webby

et al. (chapter 2) and calibrated in million years (m.y.) (Cooper 1999b). Altogether, the data compiled by the “Chitinozoan Clade Team” of IGCP project no. 410 is based on more than 10,000 productive samples. The respective position of FADs and LADs of most of the chitinozoan species can be precisely documented through fairly continuous sequences. A

295

296

   . calculation method close to that suggested by Cooper (chapter 4) for his normalized diversity measure (dnorm ) has been adopted. However, in our diversity measure, here referred to as the balanced total diversity measure (BTD), a full score is given to species confined to the time slice instead of a half score as adopted in Cooper’s dnorm . For the sake of homogeneity, the following rules have been used in counting at species level. • For the global analysis, only validly published species have been used. For regional analyses, validly published species, as well as published specimens left in open nomenclature (e.g., from theses and unpublished reports), have been taken into account. • A full score (i.e., equal to 1) is given to species ranging through a whole time slice, or, in case of very short ranging species, to species restricted to part of a time slice. • A full score (i.e., equal to 1) is also given for a species in each time slice between its FAD and LAD, even if not continuously recorded. This approach may introduce an overrepresentation of some long-ranging species resulting from poor diagnostic morphology (e.g., Conochitina chydaea and Rhabdochitina magna) or of poorly defined taxa (e.g., Lagenochitina esthonica). • A half score (i.e., equal to 0.5) is given to species having their FAD or their LAD within a time slice but extending into the succeeding or preceding time slice.

Turnover events may be documented in different ways, either by taking into account or by ignoring the duration of the time slices in m.y. • The origination (od ) and extinction (ed ) rates are calculated for each time slice by dividing the number of originations or extinctions by the total number of species recorded in the corresponding time slice. • The rate of species origination (odi ) or extinction (edi ) per m.y. are calculated by dividing the origination or extinction rates by the duration of the corresponding time slice expressed in m.y. These calculations correspond to the per capita rate of origination or of extinction suggested by Cooper (chapter 4). • The turnover ratio (TR) has been calculated for some paleoplates; it corresponds to the number of originations, minus the number of extinctions, that is, net increase/decrease (o − e), divided by the BTD recorded in a given time slice. Extinction events have negative values, whereas radiation events correspond to positive values. A value close to 0 indicates that the number of originations and the number of extinctions are more or less identical. This is frequently the case for short-ranging taxa having their LAD and FAD within

the same time slice. The TR emphasizes the major events but is less significant in the case of reduced diversity (e.g., in the Early Ordovician). ■

Regional Biodiversity North Gondwana (FP)

North Gondwana includes northern Africa, the Middle East, and the southern part of Europe. Most of the Ordovician sequences investigated in these regions (Paris 1998 and references therein) correspond to nearshore to outer-shelf environments at high or very high latitudes during most of the Ordovician. However, because of the greenhouse condition (Berner 1990) that prevailed from the Tremadocian to the middle part of the Ashgill, no permanent ice was reported in these areas until the late Ashgill (time slice TS.6c), when the major Hirnantian glaciation deeply affected all North Gondwanan environments (Brenchley and Marshall 1999 and references therein). Because of its erosive action (discontinuous record and recycling of older taxa), this glaciation greatly affected the records of Ordovician chitinozoans in North Gondwana (Paris et al. 1995; Paris et al. 2000a). Database The available data are from 139 published papers dealing with 1,415 productive samples. Additional information exists in numerous unpublished sources. More than 10,000 samples have yielded chitinozoans in North Gondwana. The data are contained mainly in oil company reports from different parts of southern Europe, North Africa, and Arabia (figure 28.1). The many unpublished literature sources are listed by Paris (1981, 1996) for southern Europe, ElaouadDebbaj (1988a) for Morocco, Oulebsir and Paris (1995) for Algeria, Paris (1988) for Libya, and Paris et al. (2000b) for Saudi Arabia. Chitinozoan data have also been compiled by Paris et al. (1998) for Mauritania. Recent investigations have concentrated on the Upper Ordovician in order to document the Hirnantian glaciation (Oulebsir and Paris 1995; Paris et al. 2000a; Bourahrouh unpubl. data) and the Middle Ordovician, where the lithology (offshore shaley facies) is usually favorable for chitinozoan preservation and recovery.

Chitinozoans FIGURE 28.2. Biodiversity of Ordovician chitinozoan species (excluding taxa in open nomenclature) from North Gondwana (compiled by F. Paris). BTD curve: balanced total diversity; TR curve: turnover ratio; od curve: species origination rate; odi : species origination rate per m.y.; ed curve: species extinction rate; edi : species extinction rate per m.y. For time slices and correlations adopted herein, see chapter 2.

The stratigraphic range of all specimens in the database is first documented in terms of the North Gondwanan chitinozoan biozones defined by Paris (1990, 1999) and then correlated with the 19 time slices of the Ordovician timescale following the calibration of Webby et al. (chapter 2). Taxonomic Diversity The BTD obtained for North Gondwanan chitinozoans (figure 28.2) shows average values ranging between 10 and 20 species per time slice (excluding specimens in open nomenclature). The BTD starts with very low values in the Tremadocian, where the clade originates, but ends with a dramatic drop in TS.6c (BTD = 6.5). The curve shows three main pos-

itive peaks during the late Early Ordovician (TS.2c), the late Mid Ordovician (TS.4c), and the middle part of the Late Ordovician (TS.5d), respectively. The most significant peak is in the upper part of the Darriwilian (BTD of 43 in TS.4c). Conversely, in addition to the major fall of diversity in TS.6c, two other significant falls of diversity are documented in the early Mid Ordovician (TS.3a–b) and in the Late Ordovician (TS.5b–c). It must be stressed that the diversity curve (BTD) of the North Gondwanan chitinozoans mirrors the curve of the available data, except in the Ashgill (figure 28.2). The highest diversity (43 species) in the late Darriwilian (TS.4c) corresponds to one of the most intensively investigated intervals (more than 200 samples). This peak can be explained by several cumulative

297

298

   . factors: (1) the duration of the time slices (3 m.y.); (2) the large variety of environments and localities investigated; (3) the thickness of the shaley sequences; and (4) the great chitinozoan diversity recorded in the Aquitaine Basin and Turkey. Both regions were located at more intermediate latitude than the Saharan localities and probably benefited from faunal transfers from Baltica (e.g., occurrence of Armoricochitina granulata in Turkey). The Late Ordovician peak in TS.5d (more than 20 species) is mainly due to the diverse assemblages recorded in Saudi Arabia (Al-Hajri 1995; Paris et al. 2000b) and in Morocco (Elaouad-Debbaj 1984, 1986; Bourahrouh unpubl. data) that filled the data gap of previous reviews (Paris 1996). Of the three recorded decreases in diversity (TS.3a–b, 5b–c, and 6b–c), the third one (BTD of 6.5 in TS.6c) is the most dramatic but not the most abrupt, as it began in TS.6a. This first-order event is of particular interest because it is contemporaneous with the Hirnantian glaciation (Brenchley and Marshall 1999) and cannot be related to the number of yielding samples (more than 300) (figure 28.2). The occurrence of reworked taxa in the Hirnantian diamictites (TS.6c) attributed to glacial processes (Oulebsir and Paris 1995) indicates that the actual diversity in the late Ashgill is even lower. The Hirnantian glaciation was finally responsible for the much reduced chitinozoan population (Paris et al. 2000a) after the steady decline from the late Caradoc diversity peak (TS.5d). Turnover The greatest diversity peak seen in the BTD curve at TS.4c is not matched in the TR curve (figure 28.2). Several positive values indicating a diversification trend (i.e., originations supplanting the extinctions) are, however, registered. The positive TR values recorded in the Early Ordovician document the radiation event. On the other hand, the positive values recorded in TS.5a and TS.5c are probably exaggerated because of a stratigraphic gap in the upper part of TS.5b and the lower part of TS.5c. The negative values for TS.1c and TS.2b may be biased because of the limited available data. The signal registered in the latest Ordovician is based on numerous assemblages from localities representing various environmental settings

(i.e., deep-water deposits as well as nearshore environments from a large range of paleolatitudes). It corresponds to a first-order event related to the Hirnantian glaciation. The origination (od ) and extinction (ed ) rates, illustrated by the number of FADs or LADs per time slice and by the odi and edi curves (figure 28.2), respectively, provide additional information. With the exception of the Early Ordovician, the curve of the origination rate (od ) is more or less similar to the BTD curve (peaks in TS.2c, 4c, and 5d). The odi curve, however, gives a better filtered signal with a clear Tremadocian radiation event (first-order event), a flourishing Mid Ordovician interval, and a slight amelioration in TS.6c (postglacial recovery and setting of the next Rhuddanian morphotypes with Spinachitina species). The edi curve perfectly demonstrates the dramatic extinction event that occurred during the late Ashgill glaciation (TS.6b–c pro parte). It also highlights an extinction peak registered within TS.5b. Discussion Once the chitinozoans had initially dispersed within the Ordovician oceans, and independently from the number of species recorded, originations in the North Gondwanan regions roughly counterbalanced extinctions, at least through the mid Arenig to the latest Darriwilian interval (TS.2c to 4c). The specific diversity (BTD) of the North Gondwanan Ordovician chitinozoans is fairly low (10 to 20 per time slice) when compared with those of other paleoplates. In contrast, their abundance is much higher than on other paleoplates. Significantly chitinozoans, or rather the population of “chitinozoan animals,” display distribution patterns similar to those of modern pelagic marine fauna, that is, low diversity but with high productivity in areas located at high latitudes (cold environments). Transgressions also seem to have had a positive impact on the diversity of chitinozoans as documented by the early Caradoc event (TS.5a–b). This is probably due to more open communication with Baltica, as shown by the occurrence of several typical Baltic species (e.g., Laufeldochitina stentor, Lagenochitina dalbyensis) in the North Gondwanan regions. The main transgressions registered in most of these regions, for example, in the mid Arenig (TS.2c), the early

Chitinozoans Darriwilian (TS.4a–b), and the early Late Ordovician (basal Nemagraptus gracilis transgression; TS.5a), are associated with increased chitinozoan originations (second-order feature).

East Gondwana (TWS) East Gondwana (i.e., Australia, Antarctica) was located at low latitudes during the Ordovician (figure 28.1). Data from East Gondwana include the Lower and Middle Ordovician chitinozoan assemblages from Australia, based on limited information from the Canning Basin (Western Australia: Combaz and Peniguel 1972; Winchester-Seeto et al. 2000; Winchester-Seeto unpubl.), Georgina Basin (western Queensland: Playford and Miller 1988; WinchesterSeeto unpubl.), and spot samples from the Tabita Formation (western New South Wales: WinchesterSeeto unpubl. data). The calculated diversity (BTD) of chitinozoans for each of the time slices is 3 for TS.2b, 9.5 for TS.2c–4a, 11 for TS.4b, and 16 for TS.4c. There is a progressive increase in the number of species from the Lower Ordovician to the upper part of the Darriwilian, which is consistent with the general trend identified by Paris (1999) to a global peak within the North Gondwanan chitinozoan-based jenkinsi Zone, that is, within TS.4c. The oldest samples employed in this compilation come from the upper Nambeet Formation (Canning Basin) representing TS.2b. The beginning of TS.2c is interpreted as coinciding with the first appearance of Conochitina langei (Combaz and Peniguel 1972; Winchester-Seeto unpubl.). The North American chitinozoan zonation has been adopted in this study because Australian chitinozoan faunas from the Lower and Middle Ordovician show a close resemblance to those from North America (Achab 1991; WinchesterSeeto et al. 2000). Except for very general information, there are not yet enough detailed biostratigraphic data for separation of TS.2c to 4a inclusive; thus they have been considered as one unit. TS.4b is indicated by the presence of Conochitina subcylindrica; however, the exact position of the beginning of this time slice is unknown. For this compilation, the boundary between TS.4b and TS.4c is taken at the beginning of Zone 04b established by Combaz and Peniguel (1972) and indicated by the first appearance of Cy-

athochitina hunderumensis and Belonechitina vibrissa (Winchester-Seeto et al. 2000 and unpubl. data). The youngest time slice for which there is adequate information is TS.4c. It contains the conodont E. suecicus, which correlates with the chitinozoan-based jenkinsi and turgida/subcylindrica zones from North America. However, Winchester-Seeto et al. (2000) found that uppermost Goldwyer and Nita formations can be correlated with the currently undefined chitinozoan zone above the Cyathochitina jenkinsi Zone in the North American zonation (Achab 1989), that is, still within TS.4c.

West Gondwana (YG) South America was mainly located in medium to high latitudes during the Ordovician. It represents the western part of Gondwana (figure 28.1). Ordovician chitinozoans have been described from northwestern Argentina (Volkheimer et al. 1980; Ottone et al. 1982, 2001), southern Bolivia (Heuse et al. 1999), and northern Brazil (Grahn 1992; Grahn and Paris 1992). The chitinozoan occurrences in Argentina and Bolivia are calibrated with the graptolite zonation and in Argentina also partly with conodont zones. In Brazil, only poorly known acritarchs are available for independent stratigraphic control. Except for the Arenig, chitinozoans from Argentina display a Laurentian affinity (Achab 1989); those from Bolivia and Brazil show a Gondwanan affinity (Paris 1990), although Laurentian elements are common in the Ashgill. Ordovician chitinozoan biodiversity of South America is low with respect to the data provided in the global curve presented by Paris (1999). The oldest faunas are from the late Tremadocian (TS.1b–d) of southern Bolivia and contain only one species, Desmochitina sp. gr. minor. An early Arenig fauna (TS.2a) comprises three species, including Conochitina decipiens. A somewhat younger chitinozoan assemblage with three species, including the index species Eremochitina brevis, has been described from TS.2c in northwestern Argentina. The Darriwilian is rather well represented. Four species, including Calpichitina cf. C. lata and Lagenochitina langei, are reported in northwestern Argentina. In northern Brazil, seven species, including Conochitina aff. C. havliceki, Lagenochitina obeligis, and Conochitina decipiens, occur

299

300

   . in equivalent TS.4a–b. The youngest Darriwilian (TS.4c) has yielded eight species, among them Calpichitina cf. C. lata, Cyathochitina jenkinsi, and Lagenochitina langei. Increasing chitinozoan diversity in the Darriwilian reflects a major sea level rise. A main flooding event in the early Caradoc is evident in the TS.5a–b interval, where eight species are known, notably Kalochitina multispinata, Belonechitina cactacea, and B. robusta. The youngest Ordovician chitinozoan assemblage from South America has been described in a stratigraphic level corresponding to a worldwide rise in sea level during TS.6b (Ashgill). Of the seven chitinozoan species that are represented in TS.6a and lower TS.6b, the most important are Armoricochitina nigerica, Lagenochitina prussica, and Tanuchitina anticostiensis.

Baltica (JN) Baltica drifted from high-intermediate to low latitudes during the Ordovician. Detailed investigations of chitinozoan assemblages show a rapid evolution in different parts of the Ordovician basin of the East European Platform. These investigations have been successful despite complications due to the presence of five main composite environmental belts in the Baltic regions (Nõlvak and Grahn 1993 and references therein). The belts are characterized by rather constant litho- and biofacies that have facilitated the construction of a detailed zonation. The difficulties encountered are due to secondary dolomitization; the occasional presence of barren marine red beds, carbonate mounds, or calcareous sandstone beds; and the limited outcrop area of Ordovician rocks in northern Estonia, Sweden (Grahn 1980, 1981), and Norway (Grahn et al. 1994). Most of the data were derived from core material in Ukraine, Belarus, Poland, Lithuania, Latvia, and Russia (St. Petersburg region), Gotland (Grahn 1982), and western and southern Estonia. The preservation of the chitinozoans varies from excellent to good in the dominantly bedded limestones (overall thicknesses commonly less than 250 m) but only acceptable to poor in graptolite shales (in Scania) or in other beds affected by thermal heating (e.g., Oslo area of Norway and central Sweden). The chitinozoan abundance usually ranges from a few to several tens of specimens per gram of rock.

Database A total of 5,362 samples have been used to construct the biodiversity curve. About 5 percent were unproductive, with some exceptions, for example, in Öland, where 78 percent of the samples from TS.1b–4b (Grahn 1980) proved to be unproductive. In general, there is a great similarity in the regional faunal logs and in the origination and extinction rates. Regional differences do not exceed 10 percent of all taxa within the same time slice interval. This constancy allowed the construction of a very useful zonation (Nõlvak and Grahn 1993; Nõlvak 1999). As explained in the introduction of this chapter, the biodiversity record was established using the LAD and the FAD of each species even in the case of some species of Cyathochitina and Conochitina, which were interpreted as Lazarus taxa in East Baltic sections (e.g., with discontinuous occurrences due to changing environments). The Ordovician chitinozoan list comprises 157 species belonging to 26 genera. The number of species per genus varies greatly. Five genera (Lagenochitina, Cyathochitina, Desmochitina, Conochitina, and Belonechitina) are represented by a relatively high number of species (15 to 24), whereas only 9 species are recorded in the genus Spinachitina and fewer than 6 species in each of the remaining genera. A number of taxa identified at generic level only have been omitted. However, new undescribed species have been included in the local database. In some Upper Ordovician portions of the East Baltic sections it will be possible to arrive at more detailed subdivision of the suggested time slices (e.g., see Kaljo et al. 1996 for TS.4c and TS.5c). In the Lower Ordovician the situation is quite the opposite: because of wellknown stratigraphic gaps in all investigated regions, the TS.1c–2a and TS.2b–c intervals are difficult to separate. Taxonomic Diversity and TR After the appearance of chitinozoans (Lagenochitina destombesi, L. esthonica) in the Tremadocian (TS.1b) of northern Estonia and Latvia, the most intensive origination and diversification took place at the beginning of the Darriwilian (TS.4a: Baltoscandian late Volkhov time; see figure 28.3). Below this level in the Lower Ordovician (TS.1b–3b), chitinozoans were

Chitinozoans FIGURE 28.3. Biodiversity of Ordovician chitinozoan species (excluding taxa in open nomenclature) from Baltica (compiled by J. Nõlvak). For diversity measure abbreviations and time-slice correlations, see figure 28.2.

recovered, with rare exceptions, from the condensed sections of northern Estonia. Chitinozoans here are short-ranging forms with Lagenochitina being dominant. In the other regions, condensed beds corresponding to this interval have turned out to be barren because of preservational factors. The BTD curve is simple and regular with low values in the Early Ordovician, followed by a noticeable rise in the Mid Ordovician and drop in the Late Ordovician. The number of investigated samples (figure 28.3) has no direct bearing on the sudden rise of the diversity curves in TS.4a, where all curves display positive peaks. The beginning of the Darriwilian was a time of rapid changes in chitinozoan assemblages, which is characterized by a relatively large number of

short-ranging species and the extinction rate reaching its maximum value. The higher BTD curve shows stable (TS.4b) and rising values reaching a maximum during TS.4c to 5a intervals. The highest diversity is recorded during the latest Darriwilian in TS.4c with 35 species, including a high number of short-ranging species, yet the origination rate curve shows a slight decline. Origination and specific diversity was relatively stable in the mid Caradoc (TS.5c, or late Viru). However, in sections studied in detail, such as the Rapla drill core in northern Estonia, it has been clearly possible to subdivide the TS.5c into three local stages (Keila, Oandu, Rakvere). The middle part of this interval exhibits a minimum diversity level of chitinozoan assemblages, which was caused by a substantial

301

302

   . extinction event in latest Keila time (63 percent of the species; see Kaljo et al. 1996). It is too early to determine whether this event in the northern Estonian sections is due to local environmental conditions (see discussion in Meidla et al. 1999). It is not registered in the more generalized curves presented herein because it is a short event counterbalanced by the higher diversity recorded in the remaining part of TS.5c. The decline in diversity and the rate of originations (od ) in the lower Ashgill (TS.6a) is attributed in part to the widely distributed red beds in Baltica, except for northern Estonia, northern Gotland, and Lithuania. In TS.6b, origination underwent a short intensive period followed by a sharp drop (latest Ordovician crisis), though the decline had already started near the Caradoc-Ashgill boundary in TS.5d. In TS.6c chitinozoans suffered major extinction (first-order event) at the upper boundary of the Spinachitina taugourdeaui Zone (see chapter 2), where one new species (Conochitina scabra) appears to continue, along with only three other species, into the Silurian. This level marks the onset of the environmental changes attributed to the Hirnantian glaciation (Kaljo et al. 2001). Interpretation The TR curve shows two main radiation events (TS.3a–b and 4b–c). The earlier event is somewhat overestimated because of the scarcity of chitinozoans in this interval. The causes are the paucity of localities suitable for chitinozoan preservation, long-ranging species, and the lack of definite extinctions. The TR negative values in TS.2b–c are biased by the lack of data. Comparison of the specific diversity curve with the number of samples curve shows no direct correlation: chitinozoan diversity in the Darriwilian and the Caradoc is relatively independent of the number of investigated samples. The origination (odi ) and extinction (edi ) curves show a good correlation, with regular changes in chitinozoan associations. They form the basis for the more detailed Darriwilian and Upper Ordovician biostratigraphy of Baltica.

Laurentia (AA, EA) Following the breakup of Rodinia and the formation of Gondwana, Laurentia (the ancestral North American continent) developed as the second largest

continental paleoplate. During the Ordovician, Laurentia roughly straddled the equator, and almost all the North American continent was submerged. Ordovician chitinozoans are known from shelf and slope deposits located at the Laurentian margin now approximately corresponding to eastern, southeastern, and western North America, the Canadian Arctic, Greenland, and Spitsbergen (see Achab 1989 and references therein). Database Laurentian chitinozoans have been described from 1,391 samples derived from 113 outcrops and drill cores. Most samples came from localities in the St. Lawrence Platform (Quebec and Ontario) and the northern Appalachians regions of eastern Canada (references in Achab 1989; Williams et al. 1999; Soufiane and Achab 2000a; and Albani et al. 2001); others are from 11 localities in the Arctic Platform (Achab and Asselin 1995; Soufiane and Achab 2000b), 22 localities in Tennessee and Alabama, the Black River Valley of New York, the Cincinnati region, and the Arbuckle Mountains of Oklahoma (Grahn and Miller 1986; Hart 1986; references in Achab 1989 and Grahn and Nöhr-Hansen 1989). A few localities have also been documented from the Interior Platform of western Canada (2), central Nevada (2), north of Greenland (1), and Spitsbergen (1) (references in Grahn and Nöhr-Hansen 1989; Martin 1992; Soufiane and Achab 2000b). As shown in figure 28.4, the Upper Ordovician is the most investigated part of the system, with more than 83 percent of the samples belonging to the Caradoc-Ashgill (TS.5a–6c) interval. More than 130 papers and abstracts record Laurentian chitinozoans, but only 30 percent provide useful information on chitinozoan assemblages and their occurrence. The database contains 303 species (137 in open nomenclature) assigned to 26 genera. Methodology All published and documented Laurentian chitinozoan species were entered in the database under their original names and their stratigraphic occurrence expressed in terms of graptolite, conodont, or chitinozoan zones, as reported by the original authors. Although only original data have been used, the fol-

Chitinozoans FIGURE 28.4. Biodiversity of Ordovician chitinozoan species (including taxa in open nomenclature) from Laurentia (compiled by A. Achab and E. Asselin). For diversity measure abbreviations and timeslice correlations, see figure 28.2.

lowing taxonomic modifications have been made: (1) the subspecies and varieties have been raised to species rank, and (2) the known occurrences or stratigraphic ranges of the species in the database were correlated with the 19 time slices of the timescale used in the IGCP 410 project (see chapter 2). A total stratigraphic range has been established for each species. We assumed that long-ranging species were continuously present between their FAD and LAD. This approach appears to be valid when the gap is small, but where larger gaps occur, for example, in species such as Conochitina chydaea, Conochitina subcylindrica, Cyathochitina hyalophrys, Lagenochitina baltica, Desmochitina complanata, and Tanuchitina

bergstroemi, it is possible that misidentifications of these species have been made. Taxonomic Diversity The BTD curve (figure 28.4) clearly shows the first documented occurrence of Laurentian chitinozoans in TS.2a at the base of the Arenig. It also shows that the diversity remains more or less constant during most of the Arenig (TS.2a–3a interval), with values ranging from 15 to 20 species. A slight decrease (about 10 species) is observed in the late Arenig (TS.3b–4a) followed by a continuous increase (11 to 40 species) during the Darriwilian (TS.4a–c). After a second decrease (31 species) in the early Caradoc

303

304

   . (TS.5a), the curve expands abruptly in the late Caradoc (TS.5c), where the highest diversity (BTD = 100) is observed. From this point, except in TS.6b, it continuously declines through the latest Caradoc and the Ashgill (TS.5d to 6c) but retains overall higher values than in the Arenig. Turnover Chitinozoan species origination or extinction rates (odi or edi ) for the various Ordovician time slices in Laurentia are shown in figure 28.4. The first occurrence of chitinozoans in the early Arenig (TS.2a) is reflected in the origination rate curve (odi = 0.5), which then fluctuates with values ranging from 0.00 to 0.23 through the Arenig (TS.2b–4a). Two pulses occur in TS.4b (Darriwilian) and TS.5a (early Caradoc), with values close to 0.30. The fluctuation appears to be less important during the Caradoc-Ashgill interval, with the rate increasing slightly from 0.12 in TS.5b to 0.24 in TS.6c. The extinction rate curve (edi ) shows three distinct intervals. The first corresponds to the Arenig (TS.2a–4a), where the extinction rate fluctuates between time slices with distinct pulses in TS.3a (0.22) and TS.4a (0.27). The edi values remain rather constant (0.09 to 0.16) through the late Darriwilian and most of the Caradoc (TS.4b–5c). From the latest Caradoc (TS.5d) onward, the extinction rate (edi ) increases continuously (from 0.06 to 0.49), paralleling the declining biodiversity at the end of the Ordovician. The TR shows no obvious correlation with the number of samples and the BTD curves. It emphasizes the events documented by the odi and edi analyses by pointing to the levels where originations or extinctions are predominant. Discussion As shown in figure 28.4, the number of fossiliferous samples and the total diversity curves are very similar, suggesting that the latter has to be corrected to eliminate the bias caused by a large number of samples in some intervals. Nevertheless, despite the small number of fossiliferous samples in the Early and Mid Ordovician (TS.2a–4a), a significant biodiversity record is evident. The highest diversity is observed in the most studied time slice (TS.5c), and

a brief increase in the diversity is evident during the Ashgill (TS.6b) while, at the same time, the number of fossiliferous samples is declining and the extinction rate increasing. All these observations can be integrated with and related to the global events and to the geologic evolution of North America, as follows: 1. In the Early Ordovician, the passive margin sediments yield a chitinozoan fauna characterized by moderate biodiversity and fluctuating origination and extinction rates, reflecting the prevalent eustatic sea level variations (second-order feature). 2. An important increase in the extinction rate, combined with a low diversity and origination rate, is observed in the early Darriwilian (TS.4a). It is followed by an increase in the biodiversity (TS.4b–c) and in the origination rate (TS.4b). These changes can be correlated with the general change from a passive to an active margin setting, corresponding to the beginning of the Taconic Orogeny and the formation of the associated foreland basin. 3. The Taconic regime prevailed through the entire Caradoc. The deepening of the basin is marked by an increase in the total diversity of the chitinozoan fauna and by more or less constant extinction and origination rates. 4. From the late Caradoc (TS.5d) onward, a significant decrease in the biodiversity and a corresponding steady increase in the extinction rate may reflect the Late Ordovician ice cap development in Gondwana and the associated sea level fall (first-order feature). Changes in environmental conditions seem to have had an impact on Laurentian chitinozoan diversity since the late Caradoc.

Avalonia ( JS, JV, MV) Detailed analyses of the fossil faunas of Europe and North America led Cocks and Fortey (1982) to suggest the existence of an independent microcontinent, Avalonia, during the Ordovician, and the Tornquist Sea, separating Avalonia from the paleocontinent Baltica (figure 28.1). Avalonia had a short independent existence from the time it started to drift off North or West Gondwana in the late Early Ordovician to its collision with Baltica in the Ashgill. It follows that evaluation of the origination/extinction rates of Ava-

Chitinozoans lonian chitinozoan faunas is limited to the Mid and Late Ordovician. The boundaries of Avalonia extended from Cape Cod in Massachusetts to the Atlantic Provinces of Canada, southern Ireland and Britain, Belgium, and northern Germany. Most strata deposited on Avalonia are now deeply buried and inaccessible. Avalonian faunas are generally poorly preserved in comparison with coeval ones from Baltoscandia and North Gondwana. Accordingly, few studies deal with Avalonian chitinozoans, and all of them have concentrated on the European part of Avalonia (i.e., eastern Avalonia). Some problems have been encountered in compiling the biostratigraphic data, such as poorly figured and described taxa and the lack of well-constrained

dating of the associated sediments. In addition, some chitinozoans are poorly preserved, and a limited number of samples were available for study (316 samples for the whole Ordovician). Database For the diversity curve (BTD), we have used both published and unpublished data. All data were critically evaluated, and only the taxa determined with certainty to the specific level were included (figure 28.5). The data came from Belgium (references in Samuelsson and Verniers 2000), Wales, Shropshire ( Jenkins 1967), and the Ebbe Anticline in western Germany (Samuelsson et al. 2002a). Also included are

FIGURE 28.5. Biodiversity of Ordovician chitinozoan species from Avalonia (compiled by J. Samuelsson and including taxa in open nomenclature) and from China (compiled by X.-F. Wang and X.-H. Chen). For diversity measure abbreviations and time-slice correlations, see figure 28.2.

305

306

   . sections where recent work has demonstrated a probable Avalonian affinity of the sediments, that is, deep drillings at Rügen, northern Germany (Samuelsson et al. 2000; Samuelsson and Servais 2001; Vecoli and Samuelsson 2001), and Pomerania, northwestern Poland (Samuelsson et al. 2002b; Samuelsson unpubl.). Taxonomic Diversity The number of species recorded in Avalonian successions is low. In all, 81 species were counted, 75 of which were recorded from the time that Avalonia existed as a separate microcontinent (i.e., from Arenig to early Ashgill). For completeness, we have also included data from the remaining Ordovician interval (TS.6b–c), which were recorded in Avalonian areas after docking, when the microcontinent ceased to exist as a separate entity. As with other North Gondwanan successions, lowdiversity chitinozoan assemblages characterize the oldest units of what was to become Avalonia (figure 28.5). Among these, Lagenochitina destombesi and Eremochitina brevis are the most distinctive taxa (Samuelsson et al. 2000). When Avalonia began its existence as an independent microcontinent, a relatively rapid diversification took place during the late Darriwilian. In general, the assemblage consists of ubiquitous, North Gondwanan taxa ( Jenkins 1967; Samuelsson and Verniers 2000; Vecoli and Samuelsson 2001). Most taxa extended to the end of the Caradoc. A less conspicuous diversification event took place at the base of the Caradoc, and the majority of these taxa (a notable exception being Lagenochitina stentor) were still present in upper Caradoc rocks. The Caradoc chitinozoans are still of the ubiquitous and North Gondwanan types, but in the later part of the Caradoc a few taxa, which can be regarded as Baltoscandian, make their appearance. No extinction event follows the two late Darriwilian and early Caradoc diversification events, although a limited number of species disappear at the end of early Caradoc time (TS.5b) (figure 28.5). In the pre-Hirnantian Ashgill, Avalonia docked with Baltica (Vecoli and Samuelsson 2001), and from the beginning of the Ashgill (TS.6a), diversity declined very rapidly (edi close to 0.6 in TS.6c). One or two taxa, however, appeared for the first time in TS.6c.

Discussion The observed high chitinozoan diversity in Avalonia during the Caradoc (TS.5a–d) can be explained in terms of the quality of the sedimentary record or the paleobiogeographic evolution, especially given that most Avalonian sections hitherto investigated are of Caradoc age. Thus, the high diversity might be due to overrepresentation of sampled Caradoc units. It is also during the Darriwilian and the Caradoc that Avalonia was an independent microcontinent in the temperate Southern Hemisphere (Cocks 2001). This position might have positively contributed to the higher diversity, since the taxa occurring on North Gondwana and in Baltoscandia have also been recovered from Avalonia. The diversity patterns can be explained as the result of a free interchange of species as Avalonia drifted northward between Gondwana and Baltica. There is no firm evidence that any chitinozoans are exclusively Avalonian. In the Ashgill, although few successions have been investigated, a major faunal turnover seems to have occurred with characteristic Baltic (Conochitina scabra) and Laurentian (Hercochitina gamachiana) taxa playing the major roles.

China (XFW, XHC) During the Ordovician, South China was located at low to moderate latitudes. Chitinozoans have been found mainly in South China (Yangtze craton and its shelf-slope biofacies) (Wang et al. 1994) and in the shelf margin of Tarim (figure 28.1). Integrated bio-, sequence-, and event-stratigraphic studies suggest that 17 chitinozoan zones and 5 major chitinozoan diversification events can be recognized. Their relationship with relevant graptolite zones and conodont zones has been established to correlate these events precisely with the time slices adopted herein (see chapter 2). Taxonomic Diversity The earliest Ordovician chitinozoans appeared in TS.1b. This event is marked by the appearance of Lagenochitina destombesi and Conochitina? sp. (figure 28.5). It is associated with the transgression at the base of the Nanjinguan Formation (Wang et al.

Chitinozoans 1996). The chitinozoan diversity increases (BTD = 3) in the lower Fenxiang Formation (TS.1c) with the appearance of Conochitina cf. decipiens, C. cf. pervulgata, C. poumoti, and Cyathochitina? cf. clepsydra (Wang et al. 1996). The appearance of the Conochitina symmetrica, the worldwide index species of the symmetrica Zone, heralds an initial chitinozoan radiation in TS.1d–2a (BTD increasing from 5.5 to 7 in this interval). C. symmetrica is found with Lagenochitina cf. obeligis, L. obeligis, Euconochitina vulgaris, and Eisenackitina tongziensis (Hou and Wang 1982; Grahn and Geng 1990; Wang and Chen 1992, 1994; Chen 1994; Chen et al. 1996). These species coexist with conodonts of the Serratognathus-Paroistodus proteus Zone, indicating that the C. symmetrica Zone is correlative with the latest Tremadocian to earliest Arenig. The overlying Eremochitina baculata Zone (TS.2b) corresponds to most of the graptolite-based approximatus and fruticosus zones. A slight increase in the chitinozoan diversity is recorded in TS.2b to TS.3a (BTD increasing from 8 to 11.5). The abundance and diversity of Conochitina and the change in chitinozoan assemblages characterize this interval. At least 12 species of chitinozoans (e.g., C. raymondi, C. brevis, C. baculata, L. esthonica) occur simultaneously or successively in this interval. The assemblage of the Eremochitina brevis Zone (TS.2c) is quite similar to that of the underlying raymondi Zone (TS.2b), except for the first appearance of the index species with graptolites of part of the suecicus Zone. Conochitina langei has its FAD in the suecicus Zone (TS.2c–3a) and ranges up to the graptolite-based austrodentatus Zone in association with C. poumoti and other taxa (Chen et al. 1996). The mixture of chitinozoans from different paleobiogeographic realms is believed to be related to warmwater currents associated with transgression and the clockwise rotation and movement of the South China plate from moderate to low southern latitudes along with the northward movement of Gondwana during this time—Dobaowanian (TS.2a–c) to early Dawanian (TS.3a) (Wang and Chen 1999). The maximum chitinozoan diversity (BTD = 22) is recorded in TS.4a with no fewer than 17 chitinozoan species referred to the Cyathochitina protocalix and C. calix zones (Geng 1984; Wang and Chen 1994

and references therein; Chen et al. 1996). This is a new stage in the evolution and diversity of the species including the chitinozoan genera Cyathochitina and Rhabdochitina. The base of the C. protocalix Zone is defined by the FAD of the index species in association with Sagenachitina oblonga, Conochitina pirum, Conochitina havliceki, and graptolites of the sinodentatus Zone (= clavus Zone). The association of S. oblonga, C. pirum, and C. havliceki is also documented from the graptolite-based clavus Zone to the lower part of the austrodentatus Zone (Chen et al. 1996). The C. calix Zone characterizes the interval from the graptolite-based upper part of the clavus to the lower part of the austrodentatus zones. Most of the chitinozoans (e.g., R. usitata, R. turgida) occurring in the C. calix Zone range up to the C. jenkinsi Zone, except the 5 species originating in the underlying protocalix Zone. The jenkinsi Zone has a diversified chitinozoan assemblage with 15 documented species (Wang and Chen 1994). The base of the zone is defined by the first appearance of the index species in association with graptolites of the upper part of the austrodentatus Zone. The FAD of C. jenkinsi therefore occurs earlier in China. Its top is below the base of the Lagenochitina shizipuensis Zone (Chen and Wang 1996). In the Yangtze platform graptolites from the sinodentatus to the austrodentatus zones (TS.3b–4a, upper Dawan Formation) are considered to be contained in Transgressive System Track (TST) deposits of Darriwilian age (Wang et al. 1996). The overlying Guniutan Formation comprises highstand regressive deposits and yields conodonts of the upper part of the variabilis Zone to the lower part of the anserinus Zone and chitinozoans of the jenkinsi Zone. The abundance and diversity of the chitinozoans in this zone decrease owing to the influence of sea level changes. In TS.5a, after a period of extinction (e.g., six taxa disappeared in TS.4c) with very few originations (e.g., one FAD in TS.4c), new chitinozoan species appeared during TS.5a (BTD = 15). The event is characterized by the occurrence of chitinozoans in deep-water settings during the earliest Caradoc associated with the largest Ordovician transgressive episode across the platform areas of South China and Tarim. Chitinozoans of the Eisenackitina uter and Lagenochitina deunffi zones occur in association with

307

308

   . graptolites of the gracilis Zone in condensed sections— the deposits associated with maximum flooding (Wang et al. 1996). A similar chitinozoan assemblage has been reported from the upper Sargan Formation of the Tarim Basin (Cai 1991; Zeng et al. 1996). Following the regression, mid to late Caradoc (TS.5b–c) chitinozoan assemblages belonging to the hirsuta, tanvillensis, and robusta zones are characterized by high abundance and low diversity (BTD respectively of 13.5 and 11 in TS.5b and TS.5c). These assemblages are known from the Tarim Basin (Cai 1991; Zeng et al. 1996) and are correlated to the graptolite-based americanus and quadrimucronatus zones (Wang et al. 1996). The progressive lowering of the chitinozoan diversity registered from the Darriwilian continued up to the Ashgill (e.g., BTD = 5 in TS.6a) (figure 28.5). During the Ashgill the Yangtze region was transformed into a semi-isolated basin following a sea level drop caused by the glaciation in North and Central Gondwana. The deposition of large amounts of organic matter generated an anoxic environment, and the accompanying change of salinity brought about a decrease in the abundance and diversity of chitinozoans in the Yangtze Basin. Only a few species belonging to the Ancyrochitina merga and Tanuchitina fusiformis zones have been found in the black shales of western Hubei, Jiangsu, and Guizhou (Qian and Geng 1989). Associated graptolites suggest that they belong to TS.6b. Deposits equivalent of the Wufeng Formation are absent or, at best, incomplete in Tarim with no Ashgill chitinozoans identified.

Siberia (NS, OO) The Siberian paleoplate was located at low latitudes during the Ordovician (figure 28.1). The Ordovician chitinozoan assemblages briefly reported here are from the Siberian Platform and the Altai Mountains. Siberian Platform Chitinozoan assemblages (15 productive samples) are known from six stratigraphic levels in terrigenous sediments of Arenig to Ashgill age (Obut and Zaslavskaya 1980; Zaslavskaya 1982, 1984). Almost all the recorded species are long-ranging. At the present

stage of investigation, no regional chitinozoan zonation is available. The recorded species are difficult to correlate with the chronostratigraphic subdivisions adopted for this volume, as the Ordovician rocks from the Siberian Platform have yielded very few graptolites and most reported conodont assemblages are endemic. Gorny Altai The Altai Mountains (Gorny Altai) belong to the western part of the Altai-Sayan folded area and form the southwestern margin of the Siberian Platform. Samples were collected from terrigenous strata in the northern, northeastern, and central parts of the Gorny Altai. Several chitinozoan assemblages (17 productive samples) were described (Zaslavskaya et al. 1978; Obut and Zaslavskaya 1980; Zaslavskaya and Obut 1984). Revision of the graptolites and chitinozoans allowed recognition of four regional zones from the Arenig to the early Ashgill: (1) Conochitina raymondi, (2) Conochitina parvicolla, (3) Cyathochitina calix, and (4) Lagenochitina dalbyensis–Desmochitina lecaniella. Only the chitinozoan assemblages occurring with graptolites are mentioned here. Two additional chitinozoans are reported from the Ordovician of the Gorny Altai: (1) Laufeldochitina aff. stentor from Arenig strata in association with graptolites of the densus Zone (Zaslavskaya et al. 1978; Petrunina et al. 1984); and (2) Lagenochitina aff. deunffi from upper Arenig strata of the graptolite-based hirundo Zone (Zaslavskaya and Obut 1984). Because Laufeldochitina stentor and Lagenochitina deunffi are respectively the index species of a late Darriwilian–early Caradoc zone in Baltica (Nõlvak and Grahn 1993) and of an early Caradoc zone in North Gondwana, the two Siberian forms might yet prove to be new species. The data concerning the Ordovician chitinozoans from Siberia are still too sparse to develop a meaningful biodiversity curve. ■

Global Biodiversity

A global database was assembled from the different regional data sets. The same generic names were adopted, but the taxa in open nomenclature were omitted. Additional adjustments were necessary in documenting the total range of some species (e.g., the

Chitinozoans

FIGURE 28.6. Global biodiversity of Ordovician chitinozoan species (excluding species in open nomenclature) and major turnover. For diversity measure abbreviations and time-slice correlations, see figure 28.2.

earliest FAD and the latest LAD were used to show the total range of each species). Of the 287 species compiled in the global database, only 2 percent are distributed across all five paleocontinents, 6 percent are distributed across four paleoplates, 9 percent are represented in three paleoplates, and 19.5 percent are found in two paleoplates. The remaining species (62.7 percent) are reported only from one paleoplate. It is apparent that any temporary input from another paleoplate increased the diversity (e.g., Caradoc assemblages from Avalonia). As with the re-

gional chitinozoan diversity data, several curves have been plotted at a global scale (figure 28.6), that is, BTD graph, TR graph, and odi and edi curves. These global curves differ from their regional counterparts, suggesting that regional factors may also have been important in the diversification of chitinozoans. The global BTD curve is depicted as a continuous diversification of the group from the early Tremadocian to a first peak in the late Darriwilian (BTD = 96 in TS.4c) and then a short-lived lowering of diversity with a loss of about 25 percent of the BTD in TS.5a

309

310

   . and 5b. A second maximum of diversification followed in the late Caradoc (BTD = 101 in TS.5c) and then decreased progressively until the end of the Ordovician (BTD = 38 in TS.6c). The TR provides additional information suggesting brief pulses of extinction in the early Arenig (TS.2b) and again in the late Caradoc (TS.5d). The latter pulse is better documented because it is based on a larger number of species records (11 originations against 48 extinctions). However, the most dramatic extinction event (only 7 originations against 46 extinctions) occurs in the latest Ashgill (TS.6c). This event is well illustrated by the steep decline of the edi curve. In contrast, no significant origination event is registered by the odi curve on a global scale, except during the initial Early Ordovician radiation of the group. The various plots in figure 28.6 clearly confirm that the high biodiversity levels are not systematically synchronous with the origination events and low diversity is not always related to extinction events. Each of these parameters must be shown separately in order to depict more clearly the biodiversification patterns of the chitinozoans. ■

Conclusions

Biodiversity appraisal requires reliable identifications, and this is probably the most difficult step to achieve because taxonomic studies of chitinozoans have been so neglected recently. Previous attempts to evaluate the chitinozoan diversification (Paris 1999; Paris and Nõlvak 1999) have shown that the size of the data set for each time slice and the duration of the time slices control the shape of the biodiversity curves. Large data sets increase the probability of recording rare taxa, and long time intervals increase the cumulative effect of very short ranging species. In order to be aware of the size of the data set per time slice, a curve illustrating the number of available samples is drawn parallel with the regional diversity curve (figures 28.2, 28.3. 28.4, 28.6). Because they have roughly the same duration (i.e., 2–3 m.y.), the adopted time slices (chapter 2) avoid this duration bias. As a consequence, the odi and edi curves integrating the duration of the time slices produce curves that are less irregular than those of the origination (od ) and extinction (ed ) rates (figures 28.2, 28.3, 28.4, 28.6). Long-ranging taxa, even if they are suspect on taxonomic grounds,

and even if they require, at times, the introduction of virtual occurrences between their FADs and LADs, do not affect the graphs significantly. The differences observed between the regional and global curves highlight the existence of various levels of changes in the diversity of Ordovician chitinozoans. The first-order trends are registered in both the regional and the global curves, whereas second-order trends are recorded only regionally, in one or two paleoplates.

First-Order Features The major feature of the chitinozoan diversification during the Ordovician is the radiation beginning in the Tremadocian and developing rather regularly to the latest Darriwilian (figures 28.2–28.6). This radiation was probably driven by such intrinsic factors as a high evolutionary potential (plasticity of the genome of the “chitinozoan animals”), as revealed by the numerous innovations recorded during the early history of the group (Paris et al. 1999). The Ordovician is usually regarded as predominantly represented by a greenhouse state (Berner 1990). The radiation was probably favored by stable climatic conditions during the Early and Mid Ordovician and a large variety of niches available for occupation of pelagic organisms such as the “chitinozoan animals.” The onset of the chitinozoan diversity decline during the Late Ordovician (mainly in the Ashgill), associated with the dramatic extinction event in the Hirnantian (TS.6c), also represents a first-order change. The Hirnantian crisis is contemporaneous with a glaciation of the first magnitude (chapter 9; Brenchley and Marshall 1999 and references therein) and thus is most probably linked to a drastic fall in temperature and drop in sea level. This relationship is confirmed by the first evidence of chitinozoan recovery in the topmost Hirnantian, just after the melting of the major part of the African ice cap (Paris et al. 2000a). The explanation for the progressive global lowering of chitinozoan diversity through the Late Ordovician (from TS.5d onward) is less clear, but it could be linked to increased volcanic activity or raised carbon dioxide levels as possible causes for this global change. It was not an intrinsic factor for the chitinozoans because the group flourished again in the Sil-

Chitinozoans urian, especially in the Pridoli, reaching an even higher diversity than in the Mid or Late Ordovician (Paris and Nõlvak 1999).

Second-Order Features These features have been reported previously from different paleoplates. In eastern Laurentia, the increase in the chitinozoan diversity culminated in the middle part of the Caradoc (TS.5c) and is related to tectonic activity (e.g., the Taconic Orogeny, which was responsible for the deepening of the Appalachian foreland basin). Large marine transgressions are also believed to have increased chitinozoan diversity, probably because they favor argillaceous sedimentation in the more distal environment (i.e., better preservation and fuller registration of the chitinozoans). This is shown by the Darriwilian transgression in China and, to a lesser extent, by the mid Darriwilian (base of the Llanvirn) transgression in North Gondwana. However, the largest Ordovician transgression at the beginning of the graptolite-based gracilis Zone is not accompanied by numerous originations in China. In North Gondwana, the positive balance between origination and extinction in TS.5a is probably linked to external factors. For example, from TS.5b to 5d, Balto-

scandian taxa probably drifted in the anticlockwise oceanic gyre of the southern oceans to higher latitudes, which were not yet disrupted or modified by the docking of Avalonia with Baltica during the Ashgill. The few meteorite effects documented in the Ordovician are not likely to have had any significant effect on patterns of chitinozoan diversity (Grahn et al. 1996).

 The authors are grateful to John Riva, Barry Webby, and Theresa Winchester-Seeto for the improvement of the text and to Merrell Miller, who reviewed the manuscript. We thank Kathleen Lauzière for her support in database management and Denis Lavoie for his constructive advice (Laurentia). We also acknowledge the help of A. Ancilletta, B. Billiaert, P. De Geest, J. Vanmeirhaeghe (Avalonia), M. Mechin (Laurentia), S. Al-Hajri, A. Bourahrouh, and L. Oulebsir (Gondwana) for use of their unpublished data. The project was supported by IGCP 410, NSER Canada (AA and EA), STINT and Uppsala University ( JS), Estonian Science Foundation grant no. 4674 ( JN), and the “CRISEVOLE” project of the French CNRS (FP).

311

29

Conodonts: Lower to Middle Ordovician Record Guillermo L. Albanesi and Stig M. Bergström

C

onodonts, eel-shaped animals that were common inhabitants of Paleozoic and Triassic seas, are now interpreted to be chordates (Aldridge et al. 1993). Recent phylogenetic analyses based on morphological, biochemical, and physiological characters suggest conodonts to be the most plesiomorphic member of the group Gnathostomata (Donoghue et al. 2000). Their apatitic feeding microelements are usually well preserved, and conodonts have a fossil record whose completeness competes with that of any other animal group (Foote and Sepkoski 1999). The excellent fossil record of conodonts and their rapid evolution make them key tools for establishing highresolution biostratigraphy from the Middle Cambrian through the Triassic. Application of advances in systematics methodology to conodonts shows them to exhibit important attributes of vertebrate phylogenetic paleontology. Coupled with the virtues of micropaleontology (with a strong tradition in stratophenetics; see Gingerich 1990), this particular fossil group has now become involved in the discussions of the importance of morphology and stratigraphy in the resolution of relationships in order to work out superior phylogenetic models (Dzik 1999; Donoghue 2001). With recent advances in the understanding of the phylogeny of the group and the access to a database that in its detail is now unrivaled among Paleozoic and Triassic fossil groups, we are now in the position to achieve a

312

much improved understanding of many aspects of conodonts and the environments in which they thrived. This will assist us in a variety of phylogenetic, paleoecologic, paleobiogeographic, paleogeographic, and paleoceanographic interpretations (Purnell 2001; Sweet and Donoghue 2001). Analyses of biodiversity, viewed as the number and variability of genes, species, and communities in space and time along with resulting processes and patterns of diversification in marine ecosystems, have always been problematic and uncertain given the incomplete preservation of fossils, inadequate sampling, and arbitrary taxonomy (Sepkoski 1997; Conway Morris 1998a). Conodonts are not an exception to this generalization; however, after the pioneer attempts to reveal general patterns of diversification (Clark et al. 1981; Aldridge 1988; Sweet 1988b; Barnes et al. 1996), modern taxonomy and much more extensive occurrence records now permit detailed studies of diversity changes in particular chronostratigraphic intervals, as exemplified by the present analysis of the Lower and lower Middle Ordovician conodont diversity. ■

Stratigraphic Framework and Taxonomic Database

We analyze a database that consists of the stratigraphic ranges of Lower and lower Middle Ordovician conodonts from well-documented localities in

Conodonts Laurentia, Baltica, the Gondwanan margin in South America, China, and Australia. Sections at selected localities record a continuous succession of biostratigraphic units, which represent diverse platform and slope environments of different paleobiogeographic nature. Most published and unpublished conodont data from these localities represent recent studies. Older significant contributions whose conodont systematics needed revision were updated using conodont multielement taxonomy. Despite our efforts, many taxa previously classified sensu formae or in open nomenclature had to be excluded from the conodont database and remain as residual taxa. The database also is constrained by the number of selected localities and by the observed taxon ranges as a confident estimate of real duration of a taxon (Hayek and Bura 2001). It is important to note that, except for a few localities with biostratigraphically well-controlled information, Ordovician conodont data from Asia and Australia are incomplete and some faunas could not be revised based on the published information. Consequently, our database represents mainly regions surrounding the Iapetus Ocean (Williams et al. 1995). However, because of the broad range of environments represented by the large number of study localities in both the Midcontinent and Atlantic realms (see tables 29.1, 29.2), we believe that the present database is extensive enough to cover all the major paleoecologic environments and to serve as a reasonably reliable basis for a first assessment of the global diversity patterns exhibited by Lower and lower Middle Ordovician conodont faunas. Our database includes 430 species representing 106 genera that are present in the stratigraphic interval from the Paltodus deltifer through the Eoplacognathus suecicus zones of the Atlantic realm, that is, from the middle part of the Tremadocian Stage of the Lower Ordovician to the upper-middle part of the Middle Ordovician. The conodonts of the lowest part of the Lower Ordovician are not considered because of the incompleteness of the information available on a global scale and because of the poorly understood relations between euconodonts and paraconodonts. We start this study at the onset of a major diversification of conodont lineages that occurred just after the existence of the lowest Ordovician Iapetognathus-Cordylodus faunas (Nicoll 1992; Nicoll et al. 1999; Dubinina 2000), that is, from the manitouensis-

deltifer interval, where a stock of taxa gave rise to a main plexus of adaptive Ordovician conodont communities (Miller 1984; Sweet 1988b). This diversification of conodont genera is an early phase of the Ordovician Radiation (cf. Miller 1997c), associated with biogeographic differentiation into the Midcontinent and Atlantic conodont realms (Barnes et al. 1973; Bergström 1990; Pohler and Barnes 1990). Stratigraphic ranges of taxa in this study span parts of Boucot’s (1983) Ecologic-Evolutionary Units (EEU) 3 and 4, whose limit coincides with the North American Sauk/Tippecanoe sequence boundary, which has been proposed as a level for the base of the global Middle Ordovician Series (Webby 1998; Finney and Ethington 2000). The time intervals for global analysis of conodont faunas used herein correspond to those of previously defined biozones within these EEUs (figures 29.1, 29.2). Analyzed intervals represent six biostratigraphic intervals, which constitute single biozones in the Lower Ordovician of the Midcontinent and Atlantic faunal realms, or combined two to three successive biozones in the lower Middle Ordovician of these realms. Combinations of these global biostratigraphic intervals of 3–5 million years’ (m.y.) duration are considered equivalent to EcologicEvolutionary Subunits (EESs), biomeres, or community groups. They include a large number of timesuccessive communities that are unique at the species level (Boucot 1983, 1990). Because of correlation difficulties and the way the conodont succession has been described, it proved impractical to use the time slice (TS ) subdivisions referred to elsewhere in this book directly for our analysis, but the approximate correlation between the time slices and our biostratigraphic intervals is as follows (figures 29.1, 29.2): manitouensis-deltifer interval, late TS.1a through TS.1b to early TS.1c; deltatus-proteus interval, late TS.1c through TS.1d to early TS.2a; communiselegans interval, middle to late TS.2a and middle to late TS.2b; andinus-evae interval, late TS.2b through TS.2c; laevis-norrlandicus interval, TS.3a–b and early TS.4a; and sinuosa-suecicus interval, late TS.4a through TS.4b to early TS.4c. Biodiversity can be evaluated by describing patterns of stability and change in marine ecosystems as they are structured on long (about 10 m.y.) timescales and defined as “coordinated stasis” (Brett and Baird 1995). These patterns involve intervals of stability

313

TABLE 29.1. Recent Contributions on Lower Ordovician Conodont Records from Selected Localities Worldwide Locality Precordillera (Cuyania Terrane), western Argentina

Northwestern Argentina

Baltoscandian Region

Lithostratigraphic Unit/Section

Reference

Biostratigraphic Interval

Bergström, unpublished collection

M-D

Rossodus manitouensis

La Silla Formation

Lehnert 1995a; Albanesi et al. 1998

M-D D-P

R. manitouensis, Paltodus deltifer Paroistodus proteus

San Juan Formation Ponón Trehué Formation

Serpagli 1974; Lehnert 1995a; Albanesi et al. 1998; Lehnert et al. 1998

D-P C-E A-E

P. proteus Priondious elegans Oepikodus evae, O. intermedius

Volcancito Formation, Famatina

Albanesi et al. 2000

M-D

Paltodus deltifer

Suri Formation, Famatina

Albanesi and Astini 2000a

A-E

O. evae, O. intermedius

Santa Victoria Group, Eastern Cordillera

Rao et al. 1994; Rao and Flores 1997; Albanesi et al. 2001

M-D D-P A-E

R. manitouensis Acodus deltatus, P. proteus O. evae

Lava River Section, western Russia

Tolmacheva 2001

M-D D-P C-E A-E

P. deltifer P. proteus P. elegans O. evae

Talubäcken Section, Sweden

Bergström 1988

D-P C-E A-E

P. proteus P. elegans O. evae

Furuhäll Section, central Öland, Sweden

Bagnoli et al. 1988

M-D D-P C-E A-E

P. deltifer P. proteus P. elegans O. evae

Horns Udde sections, northern Öland, Sweden

Bagnoli and Stouge 1997

A-E

O. evae, Trapezognathus diprion, Microzarkodina n. sp. A

Brattefors, Västergötland, Sweden

Löfgren 1997

M-D

P. deltifer

Hunneberg, Västergötland, Sweden

Löfgren 1993a

M-D D-P C-E

P. deltifer P. proteus P. elegans

Siljan District, Sweden

Löfgren 1994

M-D D-P C-E A-E

P. deltifer P. proteus P. elegans O. evae

Marathon Limestone, western Texas

Izold 1993

M-D D-P C-E A-E

R. manitouensis A. deltatus P. elegans O. evae

El Paso Formation, western Texas, New Mexico

Repetski 1982; Izold 1993

M-D

R. manitouensis, Colaptoconus quadraplicatus A. deltatus Fahraeusodus marathonensis, Oepikodus communis Jumudontus gananda/Reutterodus andinus

D-P C-E A-E

North America

Zone

San Jorge Formation

Ouachita Mountains, Arkansas and Oklahoma

Repetski and Ethington 1977; Repetski and Ethington in Stone et al. 1994

M-D D-P A-E

R. manitouensis P. proteus O. evae

Green Point Formation, Cow Head Group, St. Pauls Inlet Section, Newfoundland The Ledge Point of Head Section, Cow Head Group, Newfoundland

Johnston and Barnes 1999, 2000

D-P C-E A-E

Paracordylodus gracilis P. elegans O. evae

Stouge and Bagnoli 1988

M-D D-P C-E A-E

Prioniodus gilberti Prioniodus adami, P. oepiki, P. elegans O. evae

Ji and Barnes 1994b

M-D D-P

Cordylodus angulatus Drepanoistodus nowlani, Macerodus dianae A. deltatus, Acodus? primus O. communis, Protoprioniodus simplicissimus

Watts Bight, Boat Harbour, Catoche formations, St. George Group, Newfoundland

C-E A-E

Conodonts Locality

North America

Australia

Lithostratigraphic Unit/Section

Biostratigraphic Interval

Reference

The Ibexian Series Composite Stratotype Section and adjacent strata, House-Confusion Range Area, west-central Utah

Ross et al. 1997

Ninemile Formation, Whiterock Narrows Section, Monitor Range, Nevada

M-D D-P

Zone

C-E A-E

R. manitouensis Macerodus dianae, A. deltatus – Stultodontus costatus, O. communis Reutterodus andinus

Finney and Ethington 2000

A-E

R. andinus

Little Falls and Tribes Hill formations, eastern New York

Landing et al. 1996

M-D

R. manitouensis

Cape Weber Formation, Ella Section, East Greenland

Smith 1991

M-D D-P C-E

R. manitouensis Fauna D (middle-upper part) O. communis

Manitou Formation, Colorado

Seo and Ethington 1993

M-D

R. manitouensis

Allochthonous limestones Hensleigh Siltstone, New South Wales

Webby et al. 2000

C-E

P. elegans

Note: The database used for calculated diversity values and measures was updated following most recent revisions on multielement taxonomy for every taxon.

when species associations maintain their relative abundance with little change in composition at the species level. These intervals may be bounded by short periods of ecologic and evolutionary turnover (